Search

Language model post-training is applied to refine behaviors and unlock new skills across a wide range of recent language models, but open recipes for applying these techniques lag behind proprietary ones. The underlying training data and recipes for post-training are simultaneously the most important pieces of the puzzle and the portion with the least transparency. To bridge this gap, we introduce T\"ULU 3, a family of fully-open state-of-the-art post-trained models, alongside its data, code, and training recipes, serving as a comprehensive guide for modern post-training techniques. T\"ULU 3, which builds on Llama 3.1 base models, achieves results surpassing the instruct versions of Llama 3.1, Qwen 2.5, Mistral, and even closed models such as GPT-4o-mini and Claude 3.5-Haiku. The training algorithms for our models include supervised finetuning (SFT), Direct Preference Optimization (DPO), and a novel method we call Reinforcement Learning with Verifiable Rewards (RLVR). With T\"ULU 3, we introduce a multi-task evaluation scheme for post-training recipes with development and unseen evaluations, standard benchmark implementations, and substantial decontamination of existing open datasets on said benchmarks. We conclude with analysis and discussion of training methods that did not reliably improve performance. In addition to the T\"ULU 3 model weights and demo, we release the complete recipe -- including datasets for diverse core skills, a robust toolkit for data curation and evaluation, the training code and infrastructure, and, most importantly, a detailed report for reproducing and further adapting the T\"ULU 3 approach to more domains.

In galaxy survey analysis, the observed clustering statistics do not directly match theoretical predictions but rather have been processed by a window function that arises from the survey geometry including the sky footprint, redshift-dependent background number density and systematic weights. While window convolution of the power spectrum is well studied, for the bispectrum with a larger number of degrees of freedom, it poses a significant numerical and computational challenge. In this work, we consider the effect of the survey window in the tripolar spherical harmonic decomposition of the bispectrum and lay down a formal procedure for their convolution via a series expansion of configuration-space three-point correlation functions, which was first proposed by Sugiyama et al. (2019). We then provide a linear algebra formulation of the full window convolution, where an unwindowed bispectrum model vector can be directly premultiplied by a window matrix specific to each survey geometry. To validate the pipeline, we focus on the Dark Energy Spectroscopic Instrument (DESI) Data Release 1 (DR1) luminous red galaxy (LRG) sample in the South Galactic Cap (SGC) in the redshift bin $0.4 \leqslant z \leqslant 0.6$. We first perform convergence checks on the measurement of the window function from discrete random catalogues, and then investigate the convergence of the window convolution series expansion truncated at a finite of number of terms as well as the performance of the window matrix. This work highlights the differences in window convolution between the power spectrum and bispectrum, and provides a streamlined pipeline for the latter for current surveys such as DESI and the Euclid mission., Comment: 30 pages, 12 figures, for submission to JCAP

We present a comprehensive set of harmonic resistance measurements of the dampinglike (DL) and fieldlike (FL) torques in Pt/CoFeB, Pt/Co, W/CoFeB, W/Co, and YIG/Pt bilayers complemented by measurements of the DL torque using the magneto-optical Kerr effect and calibrated by nitrogen vacancy magnetometry on the same devices. The magnon creation-annihilation magnetoresistances depend strongly on temperature and on the magnetic and transport properties of each bilayer, affecting the estimate of both the DL and FL torque. The DL torque, the most important parameter for applications, is overestimated by a factor of 2 in W/CoFeB and by one order of magnitude in YIG/Pt when not accounting for the magnonic contribution to the planar Hall resistance. We further show that the magnonic contribution can be quantified by combining measurements of the nonlinear longitudinal and transverse magnetoresistances, thus providing a reliable method to measure the spin-orbit torques in different material systems.

Charge-spin conversion phenomena such as the spin Hall effect allow for the excitation of magnons in a magnetic layer by passing an electric current in an adjacent nonmagnetic conductor. We demonstrate that this current-induced modification of the magnon density generates an additional nonlinear longitudinal and transverse magnetoresistance for every magnetoresistance that depends on the magnetization. Using harmonic measurements, we evidence that these magnon creation-annihilation magnetoresistances dominate the second harmonic longitudinal and transverse resistance of thin Y$_{3}$Fe$_{5}$O$_{12}$/Pt bilayers. Our results apply to both insulating and metallic magnetic layers, elucidating the dependence of the magnetoresistance on applied current and magnetic field for a broad variety of systems excited by spin currents.

We present two cases of Ly$\alpha$ changing-look (CL) quasars (J1306 and J1512) along with two additional candidates (J1511 and J1602), all discovered serendipitously at $z >2$ through the Dark Energy Spectroscopic Instrument (DESI) and the Sloan Digital Sky Survey (SDSS). It is the first time to capture CL events in Ly$\alpha$ at high redshift, which is crucial for understanding underlying mechanisms driving the CL phenomenon and the evolution of high-redshift quasars and galaxies. The variability of all four sources is confirmed by the significant change of amplitude in the $r$ band ($|r_{\rm DESI}-r_{\rm SDSS}| >0.5 \ \rm mag$). We find that the accretion rate in the dim state for these CL objects corresponds to a relatively low value ($\mathscr{\dot M} \approx 2\times10^{-3}$), which suggests that the inner region of the accretion disk might be in transition between the Advection Dominated Accretion Flow ($\mathscr{\dot M}<10^{-3}\sim 10^{-2}$) and the canonical accretion disk (optically thick, geometrically thin). However, unlike in C {\sc iv} CL quasars in which broad Ly$\alpha$ remained, the broad C {\sc iv} may still persist after a CL event occurs in Ly$\alpha$, making the physical origin of the CL and ionization mechanism event more puzzling and interesting.

We present three separate void catalogs created using a volume-limited sample of the DESI Year 1 Bright Galaxy Survey. We use the algorithms VoidFinder and V2 to construct void catalogs out to a redshift of z=0.24. We obtain 1,461 interior voids with VoidFinder, 420 with V2 using REVOLVER pruning, and 295 with V2 using VIDE pruning. Comparing our catalog with an overlapping SDSS void catalog, we find generally consistent void properties but significant differences in the void volume overlap, which we attribute to differences in the galaxy selection and survey masks. These catalogs are suitable for studying the variation in galaxy properties with cosmic environment and for cosmological studies., Comment: 17 pages, 6 figures

Using early data from the Dark Energy Spectroscopic Instrument (DESI) survey, we search for AGN signatures in 410,757 line-emitting galaxies. By employing the BPT emission-line ratio diagnostic diagram, we identify AGN in 75,928/296,261 ($\approx$25.6%) high-mass ($\log (M_{\star}/\rm M_{\odot}) >$ 9.5) and 2,444/114,496 ($\approx$2.1%) dwarf ($\log (M_{\star}/\rm M_{\odot}) \leq$ 9.5) galaxies. Of these AGN candidates, 4,181 sources exhibit a broad H$\alpha$ component, allowing us to estimate their BH masses via virial techniques. This study more than triples the census of dwarf AGN as well as that of intermediate-mass black hole (IMBH; $M_{\rm BH} \le 10^6~\rm M_{\odot}$) candidates, spanning a broad discovery space in stellar mass (7 $< \log (M_{\star}/\rm M_{\odot}) <$ 12) and redshift (0.001 $< \rm z <$ 0.45). The observed AGN fraction in dwarf galaxies ($\approx$2.1%) is nearly four times higher than prior estimates, primarily due to DESI's smaller fiber size, which enables the detection of lower luminosity dwarf AGN candidates. We also extend the $M_{\rm BH}$ - $M_{\star}$ scaling relation down to $\log (M_{\star}/\rm M_{\odot}) \approx$ 8.5 and $\log (M_{\rm BH}/M_{\odot}) \approx$ 4.4, with our results aligning well with previous low-redshift studies. The large statistical sample of dwarf AGN candidates from current and future DESI releases will be invaluable for enhancing our understanding of galaxy evolution at the low-mass end of the galaxy mass function., Comment: 35 pages, 22 figures, Submitted to AAS Journals, Comments are welcome

Quantitative magnetic resonance imaging (qMRI) allows images to be compared across sites and time points, which is particularly important for assessing long-term conditions or for longitudinal studies. The multiparametric mapping (MPM) protocol is used to acquire images with conventional clinical contrasts, namely PD-, T1-, and MT-weighted volumes. Through multi-echo acquisition for each contrast and variations in flip angles between PD- and T1-weighted contrasts, parameter maps, such as proton density (PD), longitudinal relaxation rate (R1), apparent transverse relaxation rate (R2$^*$), and magnetization transfer saturation (MT$_{sat}$), can be estimated. Various algorithms have been employed to estimate these parameters from the acquired volumes. This paper extends an existing maximum a posteriori approach, which uses joint total variation regularization, by transitioning from a Gaussian noise approximation to a more physically plausible model that assumes noncentral chi-distributed noise.

It has recently been realised that there are supersymmetry-preserving non-relativistic decoupling limits associated with each half-BPS object in String Theory and M-Theory. We argue that, given a $p$-brane and a $q$-brane for which there is a quarter-BPS intersecting configuration, the $p$-brane decoupling limit of the $q$-brane's worldvolume QFT is necessarily equal to the reciprocal $q$-brane decoupling limit of the $p$-brane's worldvolume theory. This is explicitly shown for the cases of the D0-D4 and D1-D3 systems. As an application of this idea, we use the reciprocal pair of limits in the M2-M5 system to describe the dynamics of interacting self-dual strings in the six-dimensional $\mathcal{N}=(2,0)$ theory. We also discuss the limits of the dual gravitational theories and, using symmetries, argue that the duality is maintained by the limits., Comment: 62 pages

Environment has long been known to impact the evolution of galaxies, but disentangling its effects from mass evolution requires careful analysis of statistically significant samples. By implementing advanced visualisation methods to test group-finding algorithms, we utilise a mass-complete sample of galaxies to z < 0.1, comprising spectroscopic redshifts from prominent surveys such as the 2dFGRS and GAMA. Our group-finding methods identify 1,413 galaxy groups made up of 8,990 galaxies, corresponding to 36% of galaxies associated with group environments. We also search for close pairs, with separations of $r_{sep}$ < 50 h$^{-1}$ kpc and $v_{sep}$ < 500 km/s, and classify them into major ($M_{sec}/M_{prim} \leq$ 0.25) and minor ($M_{sec}/M_{prim}$ > 0.25) pairs. To examine the impact of environmental factors, we employ bespoke WISE photometry to derive a star-forming main sequence relation that shows star-formation (SF) within galaxies is pre-processed as a function of group membership. Our analysis reveals that SF in galaxies is pre-processed as a function of group membership. We observe an increase in the fraction of quiescent galaxies relative to the field as group membership rises, quantified using the environmental quenching efficiency metric ($\epsilon_{env}$). Within the star-forming population, we detect pre-processing with the relative difference in specific SF rates (${\Delta} sSFR$), showing a net decrease in SF as group membership increases, particularly at larger stellar masses. Our sample of close pairs at low stellar masses shows enhanced SF efficiencies compared to the field, while at larger masses, deficiencies are evident. Our results indicate that the small-scale environments of galaxies influence SF properties, demonstrating that galaxies do not evolve in isolation over cosmic time but are shaped by complex interactions between internal dynamics and external influences., Comment: Accepted by PASA 17/10/2024. 24 pages, 22 figures, 5 tables. Abridged abstract

Reward models (RMs) have driven the state-of-the-art performance of LLMs today by enabling the integration of human feedback into the language modeling process. However, RMs are primarily trained and evaluated in English, and their capabilities in multilingual settings remain largely understudied. In this work, we conduct a systematic evaluation of several reward models in multilingual settings. We first construct the first-of-its-kind multilingual RM evaluation benchmark, M-RewardBench, consisting of 2.87k preference instances for 23 typologically diverse languages, that tests the chat, safety, reasoning, and translation capabilities of RMs. We then rigorously evaluate a wide range of reward models on M-RewardBench, offering fresh insights into their performance across diverse languages. We identify a significant gap in RMs' performances between English and non-English languages and show that RM preferences can change substantially from one language to another. We also present several findings on how different multilingual aspects impact RM performance. Specifically, we show that the performance of RMs is improved with improved translation quality. Similarly, we demonstrate that the models exhibit better performance for high-resource languages. We release M-RewardBench dataset and the codebase in this study to facilitate a better understanding of RM evaluation in multilingual settings., Comment: 16 pages, 6 figures, 10 tables. Website: https://m-rewardbench.github.io/ , Updated results with latest models. Added more author information

The Alexandrov topology affords a well-known semantics of modal necessity and possibility. This paper develops an Alexandrov topological semantics of intuitionistic propositional modal logic internally in any elementary topos. This is done by constructing interior and closure operators on the power-object associated to a given relation in the ambient topos. When the relation is an order, these operators model intuitionistic S4; when the relation is an equivalence relation, they also model the characteristic (B) axiom of classical S5. The running example of interest arises from the Branching space-time of Nuel Belnap, which is shown to induce a histories presheaf upon which can be defined an equivalence relation of being obviously undivided at a given point event. These results have some philosophical implications. For example, we study the branching space-time example in light of the indistinguishability interpretation of epistemic modal logic. We will also study several famous first-order formulas in presheaf topos semantics such as the so-called Barcan formula. We shall see, however, that one of the Barcan converses is invalidated by a simple example of non-trivial space-time branching. This invalidates a thesis of metaphysical actualism, namely, that there are no possibly existing but non-actual entities.

A major task for moderators of online spaces is norm-setting, essentially creating shared norms for user behavior in their communities. Platform design principles emphasize the importance of highlighting norm-adhering examples and explicitly stating community norms. However, norms and values vary between communities and go beyond content-level attributes, making it challenging for platforms and researchers to provide automated ways to identify desirable behavior to be highlighted. Current automated approaches of detecting desirability are limited to measures of prosocial behavior, but we do not know whether these measures fully capture the spectrum of what communities value. In this paper, we use upvotes, which express community approval, as a proxy for desirability and conduct an analysis of highly-upvoted comments across 85 popular sub-communities on Reddit. Using a large language model, we extract values from these comments and compile 97 $\textit{macro}$, $\textit{meso}$, and $\textit{micro}$ values based on their frequency across communities. Furthermore, we find that existing computational models for measuring prosociality were inadequate to capture 86 of the values we extracted. Finally, we show that our approach can not only extract most of the qualitatively-identified values from prior taxonomies, but also uncover new values that are actually encouraged in practice. This work has implications for improving moderator understanding of their community values, motivates the need for nuanced models of desirability beyond prosocial measures, and provides a framework that can supplement qualitative work with larger-scale content analyses., Comment: Preprint: 11 pages, 3 figures, 1 table

We measure the growth of cosmic density fluctuations on large scales and across the redshift range $0.3

The Large Magellanic Cloud (LMC) is a Milky Way (MW) satellite that is massive enough to gravitationally attract the MW disc and inner halo, causing significant motion of the inner MW with respect to the outer halo. In this work, we probe this interaction by constructing a sample of 9,866 blue horizontal branch (BHB) stars with radial velocities from the DESI spectroscopic survey out to 120 kpc from the Galactic centre. This is the largest spectroscopic set of BHB stars in the literature to date, and it contains four times more stars with Galactocentric distances beyond 50 kpc than previous BHB catalogues. Using the DESI BHB sample combined with SDSS BHBs, we measure the bulk radial velocity of stars in the outer halo and observe that the velocity in the Southern Galactic hemisphere is different by 3.7$\sigma$ from the North. Modelling the projected velocity field shows that its dipole component is directed at a point 22 degrees away from the LMC along its orbit, which we interpret as the travel direction of the inner MW. The velocity field includes a monopole term that is -24 km/s, which we refer to as compression velocity. This velocity is significantly larger than predicted by the current models of the MW and LMC interaction. This work uses DESI data from its first two years of observations, but we expect that with upcoming DESI data releases, the sample of BHB stars will increase and our ability to measure the MW-LMC interaction will improve significantly., Comment: 22 pages, 19 figures. Submitted to MNRAS

In this paper, we investigate the possibility of selecting high-redshift Lyman-Break Galaxies (LBG) using current and future broadband wide photometric surveys, such as UNIONS or the Vera C. Rubin LSST, using a Random Forest algorithm. This work is conducted in the context of future large-scale structure spectroscopic surveys like DESI-II, the next phase of the Dark Energy Spectroscopic Instrument (DESI), which will start around 2029. We use deep imaging data from HSC and CLAUDS on the COSMOS and XMM-LSS fields. To predict the selection performance of LBGs with image quality similar to UNIONS, we degrade the $u, g, r, i$ and $z$ bands to UNIONS depth. The Random Forest algorithm is trained with the $u,g,r,i$ and $z$ bands to classify LBGs in the $2.5 < z < 3.5$ range. We find that fixing a target density budget of $1,100$ deg$^{-2}$, the Random Forest approach gives a density of $z>2$ targets of $873$ deg$^{-2}$, and a density of $493$ deg$^{-2}$ of confirmed LBGs after spectroscopic confirmation with DESI. This UNIONS-like selection was tested in a dedicated spectroscopic observation campaign of 1,000 targets with DESI on the COSMOS field, providing a safe spectroscopic sample with a mean redshift of 3. This sample is used to derive forecasts for DESI-II, assuming a sky coverage of 5,000 deg$^2$. We predict uncertainties on Alcock-Paczynski parameters $\alpha_\perp$ and $\alpha_{\parallel}$ to be 0.7$\%$ and 1$\%$ for $2.6

We study the [OII] profiles of emission line galaxies (ELGs) from the Early Data Release of the Dark Energy Spectroscopic Instrument (DESI). To this end, we decompose and classify the shape of [OII] profiles with the first two eigenspectra derived from Principal Component Analysis. Our results show that DESI ELGs have diverse line profiles which can be categorized into three main types: (1) narrow lines with a median width of ~50 km/s, (2) broad lines with a median width of ~80 km/s, and (3) two-redshift systems with a median velocity separation of ~150 km/s, i.e., double-peak galaxies. To investigate the connections between the line profiles and galaxy properties, we utilize the information from the COSMOS dataset and compare the properties of ELGs, including star-formation rate (SFR) and galaxy morphology, with the average properties of reference star-forming galaxies with similar stellar mass, sizes, and redshifts. Our findings show that on average, DESI ELGs have higher SFR and more asymmetrical/disturbed morphology than the reference galaxies. Moreover, we uncover a relationship between the line profiles, the excess SFR and the excess asymmetry parameter, showing that DESI ELGs with broader [OII] line profiles have more disturbed morphology and higher SFR than the reference star-forming galaxies. Finally, we discuss possible physical mechanisms giving rise to the observed relationship and the implications of our findings on the galaxy clustering measurements, including the halo occupation distribution modeling of DESI ELGs and the observed excess velocity dispersion of the satellite ELGs., Comment: 24 pages, 13 figures, submitted to ApJ

We investigate the potential of deliberation to create consensus among fully-informed citizens. Our approach relies on two cognitive assumptions: i. citizens need a thinking frame (or perspective) to consider an issue; and ii. citizens cannot consider all relevant perspectives simultaneously: they are incompatible in the mind. These assumptions imply that opinions are intrinsically contextual. Formally, we capture contextuality in a simple quantum-like cognitive model. We consider a binary voting problem, in which two citizens with incompatible thinking frames and initially opposite voting intentions deliberate under the guidance of a benevolent facilitator. We find that when citizens consider alternative perspectives, their opinion may change. When the citizens' perspectives are two-dimensional and maximally uncorrelated, the probability for consensus after two rounds of deliberation reaches 75%; and this probability increases proportionally with the dimensionality (namely, the richness) of the perspectives. When dealing with a population of citizens, we also elaborate a novel rationale for working in subgroups. The contextuality approach delivers a number of insights. First, the diversity of perspectives is beneficial and even necessary for deliberations to overcome initial disagreement. Second, successful deliberation demand the active participation of citizens in terms of "putting themselves in the other's shoes". Third, well-designed procedures managed by a facilitator are necessary to secure increased probability for consensus. A last insight is that the richness of citizens' thinking frames is beneficial, while the optimal strategy entails focusing deliberation on a properly reduced problem., Comment: 28+2 pages, 3 figures

We investigate non-Markovian transport dynamics and signatures of the Kondo effect in a single impurity Anderson model. The model consists of a quantum dot (QD) with ultra-strong coupling to a left lead and weak coupling to a right lead acting as a detector. We calculate the waiting time distribution (WTD) of electrons tunneling into the detector using a combination of the hierarchical equations of motion approach (HEOM) and a dressed master equation. Oscillations emerge in the short-time WTD, becoming more pronounced with stronger left-lead coupling. Fourier analysis reveals a blue shift in the oscillation frequency as coupling increases, indicating enhanced system-bath hybridization. Crucially, comparison with a dressed master equation confirms that these oscillations are a direct consequence of non-Markovian system-bath correlations. We examine the Kondo effect's influence on these oscillations by varying the quantum dot's Coulomb repulsion. Increasing this interaction enhances the WTD oscillations, coinciding with the signatures of a strengthened Kondo resonance in the quantum dot's density of states. Our results demonstrate that WTD oscillations offer a valuable tool for probing non-Markovian system-bath interactions and the emergence of Kondo correlations within quantum dot systems., Comment: 14 pages, 12 figures

The highly conducting and transparent inorganic perovskites SrBO$_3$ with V, Nb, Mo, and their mixtures at the B-site have recently attracted the attention of the oxide electronics community as novel alternative transparent conducting oxides. For different applications from solar cells to transparent electronics, it is desirable to tune the optical transmission window in the ultraviolet (UV), visible and infrared (IR) range. The conventional approach is substitutional design at the A- and/or B-site. Here, we suggest a method by engineering the off-stoichiometry of the perovskite, opening new pathways to broaden the range of applications without adding additional elements. For oxide molecular beam epitaxy grown SrNbO$_3$ on GdScO$_3$ substrates, we show that controlled Sr deficiency shifts the plasma edge from about 2 eV in the visible range into the near-infrared region, 1.37 eV (similar to stoichiometric SrVO$_3$). Here, epitaxial growth allows going beyond the limitations of phase stability set by thermodynamics. The suggested approach opens a new design toolbox by including controlled vacancy sites as quasi-substitutional virtual elements., Comment: 19 pages

Molecular beam epitaxy (MBE) is a state-of-the-art technique for depositing thin films with precise stoichiometric control. However, when depositing oxides of perovskite-type ABO3, this process becomes challenging as controlling the flux rate of A and B simultaneously in the presence of oxygen is difficult. In this work, by utilizing e-beam-assisted oxide MBE, we successfully deposited SrNbO3 epitaxial thin films. A buffer layer of SrTiO3 perovskite oxide, prepared by pulsed laser deposition (PLD), was used to improve the growth of this oxide significantly. This method overcomes the ultimatum of stabilizing perovskite oxide growth using MBE since the valence mismatch between the substrate and the films were effectively masked by the ultra-thin SrTiO3 and further supported interfacial charge transfer. Interestingly, we also discovered that the perovskite oxides SrNbO3 and SrTiO3 replace K in KTaO3 substrates by Sr, regardless of the deposition technique used, which is generally considered a reason for many interfacial effects. The growth of SrNbO3 was unaffected by the K deficiency when the stable buffer layer was introduced., Comment: 17 pages

As small radiation fields play an ever-increasing role in radiation therapy, accurate dosimetry of these fields becomes critical to ensure high quality dose calculation and treatment optimization. The PRB-0002, a new plastic scintillation detector (PSD) part of the Hyperscint RP-200 dosimetric platform from Medscint, can potentially facilitate small field measurements. In this work our objective is twofold: first we perform a detailed characterization of the field output correction factors of the PRB-0002 over a wide range of field sizes; second we demonstrate how this PSD can be used to determine the field output correction factors for other small field detectors. Uncertainties were carefully considered and a detailed uncertainty budget is provided. EGSnrc Monte Carlo simulations were conducted to determine the impact of detector composition, surrounding materials, dose averaging within the sensitive volume as well as ionization quenching. From these simulations, the field output correction factors of this PSD were determined. Then, by experimental comparisons, field output correction factors for 2 solid state detectors and 3 small volume ion chambers were. With PRB-0002, the material composition was found to be balanced with the ionization quenching making the field output correction factor near unity. For fields between 0.6 $\times$ 0.6 and 30 $\times$ 30~cm$^2$, the field output correction factors of the PRB-0002 were between 1.002 and 0.999 with a total uncertainty of 0.5 %. Analysis of the uncertainty budget showed that an overall uncertainty of 0.59% can be achieved for a 1 $\times$ 1~cm$^2$ using PRB-0002. We conclude that PRB-0002 is a near-ideal detector for small field dosimetry. Furthermore, it can be used to experimentally determine the field output correction factors of other dosimeters with great accuracy.

Social media platform design often incorporates explicit signals of positive feedback. Some moderators provide positive feedback with the goal of positive reinforcement, but are often unsure of their ability to actually influence user behavior. Despite its widespread use and theory touting positive feedback as crucial for user motivation, its effect on recipients is relatively unknown. This paper examines how positive feedback impacts Reddit users and evaluates its differential effects to understand who benefits most from receiving positive feedback. Through a causal inference study of 11M posts across 4 months, we find that users who received positive feedback made more frequent (2% per day) and higher quality (57% higher score; 2% fewer removals per day) posts compared to a set of matched control users. Our findings highlight the need for platforms and communities to expand their perspective on moderation and complement punitive approaches with positive reinforcement strategies., Comment: 23 pages, 5 figures

This study reports a comprehensive environmental scan of the generative AI (GenAI) infrastructure in the national network for clinical and translational science across 36 institutions supported by the Clinical and Translational Science Award (CTSA) Program led by the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH) at the United States. With the rapid advancement of GenAI technologies, including large language models (LLMs), healthcare institutions face unprecedented opportunities and challenges. This research explores the current status of GenAI integration, focusing on stakeholder roles, governance structures, and ethical considerations by administering a survey among leaders of health institutions (i.e., representing academic medical centers and health systems) to assess the institutional readiness and approach towards GenAI adoption. Key findings indicate a diverse range of institutional strategies, with most organizations in the experimental phase of GenAI deployment. The study highlights significant variations in governance models, with a strong preference for centralized decision-making but notable gaps in workforce training and ethical oversight. Moreover, the results underscore the need for a more coordinated approach to GenAI governance, emphasizing collaboration among senior leaders, clinicians, information technology staff, and researchers. Our analysis also reveals concerns regarding GenAI bias, data security, and stakeholder trust, which must be addressed to ensure the ethical and effective implementation of GenAI technologies. This study offers valuable insights into the challenges and opportunities of GenAI integration in healthcare, providing a roadmap for institutions aiming to leverage GenAI for improved quality of care and operational efficiency.

Aims. We present an extensive catalog of the physical properties of more than a million galaxies within the Dark Energy Spectroscopic Instrument (DESI), one of the largest spectroscopic surveys to date. Spanning over a full variety of target types, including emission line galaxies and luminous red galaxies as well as quasars, our survey encompasses an unprecedented range of spectroscopic redshifts, stretching from 0 to 6. Methods. The physical properties, such as stellar masses and star formation rates, are derived via the CIGALE spectral energy distribution (SED) fitting code accounting for the contribution coming from active galactic nuclei (AGN). Based on the modeling of the optical-mid-infrared (grz complemented by WISE photometry) SEDs, we study galaxy properties with respect to their location on the main sequence. Results. We revise the dependence of stellar mass estimates on model choices and availability of the WISE photometry. The WISE information is mandatory to minimize the misclassification of star-forming galaxies as AGN. The lack of WISE bands in SED fits leads to elevated AGN fractions for 68% of star-forming galaxies identified using emission line diagnostic diagram but does not significantly affect their stellar mass nor star formation estimates., Comment: resubmitted after addressing minor referee comments

During infectious disease outbreaks, estimates of time-varying pathogen transmissibility, such as the instantaneous reproduction number R(t) or epidemic growth rate r(t), are used to inform decision-making by public health authorities. For directly transmitted infectious diseases, the renewal equation framework is a widely used method for measuring time-varying transmissibility. The framework uses information on the typical time elapsing between an infection and the offspring infections (quantified by the generation time distribution), and R(t), to describe the rate at which currently infected individuals generate new infections. For diseases with transmission cycles involving hosts and vectors, however, renewal equation models have been far less used. This is likely due to difficulties in mechanistically defining generation times that can capture the complexity of multi-stage, human-vector relationships. Here, using dengue as an example, we provide general renewal equations that are derived from first principles using age-structured systems of coupled partial differential equations across human and vector sub-populations. Our framework tracks the multi-stage transmission cycle over calendar time and across stage-specific ages, resulting in governing renewal equations that quantify how the rate at which new infections are generated from existing infections depends on stage-specific processes. The framework provides a foundation on which to base inferential frameworks for estimating R(t) and r(t) for infectious diseases with multiple stages in the transmission cycle, Comment: 14 pages, 3 figures

This paper introduces the hypothetical track-length fitting algorithm, a novel method for measuring the kinetic energies of ionizing particles in liquid argon time projection chambers (LArTPCs). The algorithm finds the most probable offset in track length for a track-like object by comparing the measured ionization density as a function of position with a theoretical prediction of the energy loss as a function of the energy, including models of electron recombination and detector response. The algorithm can be used to measure the energies of particles that interact before they stop, such as charged pions that are absorbed by argon nuclei. The algorithm's energy measurement resolutions and fractional biases are presented as functions of particle kinetic energy and number of track hits using samples of stopping secondary charged pions in data collected by the ProtoDUNE-SP detector, and also in a detailed simulation. Additional studies describe impact of the dE/dx model on energy measurement performance. The method described in this paper to characterize the energy measurement performance can be repeated in any LArTPC experiment using stopping secondary charged pions.

Today's most advanced multimodal models remain proprietary. The strongest open-weight models rely heavily on synthetic data from proprietary VLMs to achieve good performance, effectively distilling these closed models into open ones. As a result, the community is still missing foundational knowledge about how to build performant VLMs from scratch. We present Molmo, a new family of VLMs that are state-of-the-art in their class of openness. Our key innovation is a novel, highly detailed image caption dataset collected entirely from human annotators using speech-based descriptions. To enable a wide array of user interactions, we also introduce a diverse dataset mixture for fine-tuning that includes in-the-wild Q&A and innovative 2D pointing data. The success of our approach relies on careful choices for the model architecture details, a well-tuned training pipeline, and, most critically, the quality of our newly collected datasets, all of which will be released. The best-in-class 72B model within the Molmo family not only outperforms others in the class of open weight and data models but also compares favorably against proprietary systems like GPT-4o, Claude 3.5, and Gemini 1.5 on both academic benchmarks and human evaluation. We will be releasing all of our model weights, captioning and fine-tuning data, and source code in the near future. Select model weights, inference code, and demo are available at https://molmo.allenai.org.

X-ray photoelectron spectroscopy (XPS) is a versatile tool for identifying the chemical and electronic state of an element. However, the presence of various initial- and final-state effects makes the interpretation of XPS spectra tricky and erroneous. Metal oxides show multipeak XPS spectra when the charge carriers are coupled with the core hole of the photoion after photoemission. The extra peaks are often misconstrued to originate from other oxidation states of the cation. With a systematic approach to the analysis of Nb 3d XPS spectra of partially oxidized NbOx thin films prepared by molecular beam epitaxy, we found that the valence band electrons close to the Fermi level contribute to a satellite peak in Nb-oxides. Insulating Nb-oxide containing Nb5+ (d0) with no or fewer electrons near the Fermi level has no such peaks; on the other hand, Nb5+ containing thin films exhibit this satellite peak when sufficient surface charge carriers contribute to the core hole-electron coupling., Comment: 18 pages including supplementary information

The paper written in 1925 by G. Udny Yule that we celebrate in this special issue introduces several novelties and results that we recall in detail. First, we discuss Yule (1925)'s main legacies over the past century, focusing on empirical frequency distributions with heavy tails and random tree models for phylogenies. We estimate the year when Yule's work was re-discovered by scientists interested in stochastic processes of population growth (1948) and the year from which it began to be cited (1951, Yule's death). We highlight overlooked aspects of Yule's work (e.g., the Yule process of Yule processes) and correct some common misattributions (e.g., the Yule tree). Second, we generalize Yule's results on the average frequency of genera of a given age and size (number of species). We show that his formula also applies to the age $A$ and size $S$ of any randomly chosen genus and that the pairs $(A_i, S_i)$ are equally distributed and independent across genera. This property extends to triples $(H_i, A_i, S_i)$, where $(H_i)$ are the coalescence times of the genus phylogeny, even when species diversification within genera follows any integer-valued process, including species extinctions. Studying $(A, S)$ in this broader context allows us to identify cases where $S$ has a power-law tail distribution, with new applications to urn schemes., Comment: 33 pages, 4 figures. To appear in theme issue of Philosophical Transactions of the Royal Society Series B on "Phylogenetic models a century beyond G. U. Yule's 'mathematical theory of evolution'"

Besides controlling wave trajectory inside complex media, wave velocity constitutes a relevant bio-marker for medical imaging. In a transmission configuration, wave-front distortions can be unscrambled to provide a map of the wave velocity landscape $c(\mathbf{r})$. However, most in-vivo applications correspond to a reflection configuration for which only back-scattered echoes generated by short-scale fluctuations of $c(\mathbf{r})$ can be harvested. Under a single scattering assumption, this speckle wave-field cannot provide the long-scale variations of $c(\mathbf{r})$. In this paper, we go beyond the first Born approximation and show how a map of $c(\mathbf{r})$ can be retrieved in epi-detection. To that aim, a reflection matrix approach of wave imaging is adopted. While standard reflection imaging methods generally rely on confocal focusing operations, matrix imaging consists in decoupling the location of the incident and received focal spots. Following this principle, a self-portrait of the focusing process can be obtained around each point of the medium. The Gouy phase shift exhibited by each focal spot is leveraged to finely monitor the wave velocity distribution $c(\mathbf{r})$ inside the medium. Experiment in a tissue-mimicking phantom and numerical simulations are first presented to validate our method. Speed-of-sound tomography is then applied to ultrasound data collected on the liver of a difficult-to-image patient. Beyond paving the way towards quantitative ultrasound, our approach can also be extremely rewarding for standard imaging. Indeed, each echo can be assigned to the actual position of a scatterer. It allows an absolute measurement of distance, an observable often used for diagnosis but generally extremely sensitive to wave velocity fluctuations., Comment: 45 pages, 9 figures, 3 tables

We introduce Playground v3 (PGv3), our latest text-to-image model that achieves state-of-the-art (SoTA) performance across multiple testing benchmarks, excels in graphic design abilities and introduces new capabilities. Unlike traditional text-to-image generative models that rely on pre-trained language models like T5 or CLIP text encoders, our approach fully integrates Large Language Models (LLMs) with a novel structure that leverages text conditions exclusively from a decoder-only LLM. Additionally, to enhance image captioning quality-we developed an in-house captioner, capable of generating captions with varying levels of detail, enriching the diversity of text structures. We also introduce a new benchmark CapsBench to evaluate detailed image captioning performance. Experimental results demonstrate that PGv3 excels in text prompt adherence, complex reasoning, and accurate text rendering. User preference studies indicate the super-human graphic design ability of our model for common design applications, such as stickers, posters, and logo designs. Furthermore, PGv3 introduces new capabilities, including precise RGB color control and robust multilingual understanding., Comment: Project page: https://playground.com/pg-v3

The Gemini South telescope is now equipped with a new high-resolution spectrograph called GHOST (the Gemini High-resolution Optical SpecTrograph). This instrument provides high-efficiency, high-resolution spectra covering 347-1060 nm in a single exposure of either one or two targets simultaneously, along with precision radial velocity spectroscopy utilizing an internal calibration source. It can operate at a spectral element resolving power of either 76000 or 56000, and can reach a SNR$\sim$5 in a 1hr exposure on a V$\sim$20.8 mag target in median site seeing, and dark skies (per resolution element). GHOST was installed on-site in June 2022, and we report performance after full integration to queue operations in November 2023, in addition to scientific results enabled by the integration observing runs. These results demonstrate the ability to observe a wide variety of bright and faint targets with high efficiency and precision. With GHOST, new avenues to explore high-resolution spectroscopy have opened up to the astronomical community. These are described, along with the planned and potential upgrades to the instrument., Comment: Accepted in the Astronomical Journal; 26 pages, 24 figures and 4 tables

On large scales, measurements of the Lyman-$\alpha$ forest offer insights into the expansion history of the Universe, while on small scales, these impose strict constraints on the growth history, the nature of dark matter, and the sum of neutrino masses. This work introduces ForestFlow, a cosmological emulator designed to bridge the gap between large- and small-scale Lyman-$\alpha$ forest analyses. Using conditional normalizing flows, ForestFlow emulates the 2 Lyman-$\alpha$ linear biases ($b_\delta$ and $b_\eta$) and 6 parameters describing small-scale deviations of the 3D flux power spectrum ($P_\mathrm{3D}$) from linear theory. These 8 parameters are modeled as a function of cosmology $\unicode{x2013}$ the small-scale amplitude and slope of the linear power spectrum $\unicode{x2013}$ and the physics of the intergalactic medium. Thus, in combination with a Boltzmann solver, ForestFlow can predict $P_\mathrm{3D}$ on arbitrarily large (linear) scales and the 1D flux power spectrum ($P_\mathrm{1D}$) $\unicode{x2013}$ the primary observable for small-scale analyses $\unicode{x2013}$ without the need for interpolation or extrapolation. Consequently, ForestFlow enables for the first time multiscale analyses. Trained on a suite of 30 fixed-and-paired cosmological hydrodynamical simulations spanning redshifts from $z=2$ to $4.5$, ForestFlow achieves $3$ and $1.5\%$ precision in describing $P_\mathrm{3D}$ and $P_\mathrm{1D}$ from linear scales to $k=5\,\mathrm{Mpc}^{-1}$ and $k_\parallel=4\,\mathrm{Mpc}^{-1}$, respectively. Thanks to its parameterization, the precision of the emulator is also similar for both ionization histories and two extensions to the $\Lambda$CDM model $\unicode{x2013}$ massive neutrinos and curvature $\unicode{x2013}$ not included in the training set. ForestFlow will be crucial for the cosmological analysis of Lyman-$\alpha$ forest measurements from the DESI survey., Comment: 17 pages, 11 figures. Submitted to A&A

We present new Galactic reddening maps of the high Galactic latitude sky using DESI imaging and spectroscopy. We directly measure the reddening of 2.6 million stars by comparing the observed stellar colors in $g-r$ and $r-z$ from DESI imaging with the synthetic colors derived from DESI spectra from the first two years of the survey. The reddening in the two colors is on average consistent with the \cite{fitzpatrick_correcting_1999} extinction curve with $R_\mathrm{V}=3.1$. We find that our reddening maps differ significantly from the commonly used \cite{schlegel_maps_1998} (SFD) reddening map (by up to 80 mmag in $E(B-V)$), and we attribute most of this difference to systematic errors in the SFD map. To validate the reddening map, we select a galaxy sample with extinction correction based on our reddening map, and this yields significantly better uniformity than the SFD extinction correction. Finally, we discuss the potential systematic errors in the DESI reddening measurements, including the photometric calibration errors that are the limiting factor on our accuracy. The $E(g-r)$ and $E(g-r)$ maps presented in this work, and for convenience their corresponding $E(B-V)$ maps with SFD calibration, are publicly available., Comment: Submitted to the Open Journal of Astrophysics. Associated data files: https://data.desi.lbl.gov/public/papers/mws/desi_dust/y2/v1/maps/

The motility of living things and synthetic self-propelled objects is often described using Active Brownian particles. To capture the interaction of these particles with their often complex environment, this model can be augmented with empirical forces or torques, for example, to describe their alignment with an obstacle or wall after a collision. Here, we assess the quality of these empirical models by comparing their output predictions with trajectories of rod-shaped active particles that scatter sterically at a flat wall. We employ a classical least-squares method to evaluate the instantaneous torque. In addition, we lay out a Bayesian inference procedure to construct the posterior distribution of plausible model parameters. In contrast to the least squares fit, the Bayesian approach does not require orientational data of the active particle and can readily be applied to experimental tracking data.

We introduce OLMoE, a fully open, state-of-the-art language model leveraging sparse Mixture-of-Experts (MoE). OLMoE-1B-7B has 7 billion (B) parameters but uses only 1B per input token. We pretrain it on 5 trillion tokens and further adapt it to create OLMoE-1B-7B-Instruct. Our models outperform all available models with similar active parameters, even surpassing larger ones like Llama2-13B-Chat and DeepSeekMoE-16B. We present various experiments on MoE training, analyze routing in our model showing high specialization, and open-source all aspects of our work: model weights, training data, code, and logs., Comment: 61 pages (24 main), 36 figures, 14 tables

Controlling quantum spins using electric rather than magnetic fields promises significant architectural advantages for developing quantum technologies. In this context, spins in molecular nanomagnets offer tunability of spin-electric couplings (SEC) by rational chemical design. Here we demonstrate systematic control of SECs in a family of Mn(II)-containing molecules via chemical engineering. The trigonal bipyramidal (tbp) molecular structure with C3 symmetry leads to a significant molecular electric dipole moment that is directly connected to its magnetic anisotropy. The interplay between these two features gives rise to significant experimentally observed SECs, which can be rationalised by wavefunction theoretical calculations. Our findings guide strategies for the development of electrically controllable molecular spin qubits for quantum technologies., Comment: 8 pages, 3 figures, 1 table

We propose a new system for automatic 2D floorplan reconstruction that is enabled by SALVe, our novel pairwise learned alignment verifier. The inputs to our system are sparsely located 360$^\circ$ panoramas, whose semantic features (windows, doors, and openings) are inferred and used to hypothesize pairwise room adjacency or overlap. SALVe initializes a pose graph, which is subsequently optimized using GTSAM. Once the room poses are computed, room layouts are inferred using HorizonNet, and the floorplan is constructed by stitching the most confident layout boundaries. We validate our system qualitatively and quantitatively as well as through ablation studies, showing that it outperforms state-of-the-art SfM systems in completeness by over 200%, without sacrificing accuracy. Our results point to the significance of our work: poses of 81% of panoramas are localized in the first 2 connected components (CCs), and 89% in the first 3 CCs. Code and models are publicly available at https://github.com/zillow/salve., Comment: Accepted at ECCV 2022