Search

We investigate size variation with rest-frame wavelength for star-forming galaxies based on the second JWST Advanced Deep Extragalactic Survey data release. Star-forming galaxies are typically smaller at longer wavelength from UV-to-NIR at $z<3.5$, especially for more massive galaxies, indicating the inside-out assembly with in-situ star formation if ignoring dust attenuation. The size variation with wavelength shows strong dependence on stellar mass, and shows little or no dependence on redshift, specific star formation rate and galaxy environment. This suggests that the size growth of star-forming galaxies is a self-regulated process primarily governed by stellar mass. We model size as a function of both mass and redshift simultaneously, obtaining $R_{\rm e} \propto M_*^{0.23} (1+z)^{-1.04}$ at a wavelength of 0.45 ${\mu \mathrm{m}}$, and $R_{\rm e} \propto M_*^{0.20} (1+z)^{-1.08}$ at 1.0 ${\mu \mathrm{m}}$. Based on this size evolution and the star formation main sequence from the literature, we obtain the locus of typical size growth for individual galaxies of different masses on the mass-size plane. The moving trend of galaxies on the mass-size plane, which indicates the slopes of their locus, strongly correlates with the size ratio between 0.45 ${\mu \mathrm{m}}$ and 1.0 ${\mu \mathrm{m}}$, supporting the idea that the size variation with wavelength provides important information on size growth of galaxies on short timescales., Comment: Accepted for publication in ApJ, 19 pages, 11 figures

We introduce Constrained Human-AI Cooperation (CHAIC), an inclusive embodied social intelligence challenge designed to test social perception and cooperation in embodied agents. In CHAIC, the goal is for an embodied agent equipped with egocentric observations to assist a human who may be operating under physical constraints -- e.g., unable to reach high places or confined to a wheelchair -- in performing common household or outdoor tasks as efficiently as possible. To achieve this, a successful helper must: (1) infer the human's intents and constraints by following the human and observing their behaviors (social perception), and (2) make a cooperative plan tailored to the human partner to solve the task as quickly as possible, working together as a team (cooperative planning). To benchmark this challenge, we create four new agents with real physical constraints and eight long-horizon tasks featuring both indoor and outdoor scenes with various constraints, emergency events, and potential risks. We benchmark planning- and learning-based baselines on the challenge and introduce a new method that leverages large language models and behavior modeling. Empirical evaluations demonstrate the effectiveness of our benchmark in enabling systematic assessment of key aspects of machine social intelligence. Our benchmark and code are publicly available at https://github.com/UMass-Foundation-Model/CHAIC., Comment: NeurIPS 2024 Dataset and Benchmark Track. The first two authors contributed equally. Project Website at https://vis-www.cs.umass.edu/CHAIC/

We analyze the properties of a rare population, the strongly bulge-dominated early-type galaxies (referred to as sBDEs) with significant HI gas, using the databases from the FAST All Sky HI survey (FASHI) and the Arecibo Legacy Fast ALFA (ALFALFA) survey. We select the sBDEs from the Sloan Digital Sky Survey (SDSS) and cross-match with the FASHI-ALFALFA combined HI sample, resulting in 104 HI-rich sBDEs. These sBDEs tend to have extremely high HI reservoirs, which is rare in previous studies such as ATLAS$^{3D}$. 70% of the selected sBDEs are classified as quiescent galaxies, even though they have a large HI reservoir. We study the properties of these sBDEs from five main aspects: stellar population, gas-phase metallicity, stacked HI spectra, environment, and spatially resolved MaNGA data. The majority of HI-rich sBDEs appear to show lower gas-phase metallicity and are located in significantly lower-density environments, suggesting an external origin for their HI gas. We find that star-forming sBDEs exhibit statistically higher star formation efficiency and slightly older stellar populations compared to normal star-forming galaxies, suggesting a recent star formation on Gyr-timescale. They also show narrower and more concentrated HI profiles compared to control star-forming galaxies, which may explain their higher star formation efficiency., Comment: 18 pages, 14 figures, 1 table. Accepted for publication in ApJ

Morphological classification conveys abundant information on the formation, evolution, and environment of galaxies. In this work, we refine the two-step galaxy morphological classification framework ({\tt\string USmorph}), which employs a combination of unsupervised machine learning (UML) and supervised machine learning (SML) techniques, along with a self-consistent and robust data preprocessing step. The updated method is applied to the galaxies with $I_{\rm mag}<25$ at $0.2

In this paper, we investigate the problem of embodied multi-agent cooperation, where decentralized agents must cooperate given only partial egocentric views of the world. To effectively plan in this setting, in contrast to learning world dynamics in a single-agent scenario, we must simulate world dynamics conditioned on an arbitrary number of agents' actions given only partial egocentric visual observations of the world. To address this issue of partial observability, we first train generative models to estimate the overall world state given partial egocentric observations. To enable accurate simulation of multiple sets of actions on this world state, we then propose to learn a compositional world model for multi-agent cooperation by factorizing the naturally composable joint actions of multiple agents and compositionally generating the video. By leveraging this compositional world model, in combination with Vision Language Models to infer the actions of other agents, we can use a tree search procedure to integrate these modules and facilitate online cooperative planning. To evaluate the efficacy of our methods, we create two challenging embodied multi-agent long-horizon cooperation tasks using the ThreeDWorld simulator and conduct experiments with 2-4 agents. The results show our compositional world model is effective and the framework enables the embodied agents to cooperate efficiently with different agents across various tasks and an arbitrary number of agents, showing the promising future of our proposed framework. More videos can be found at https://vis-www.cs.umass.edu/combo/., Comment: 23 pages. The first three authors contributed equally

Galaxy interactions and mergers can enhance or reduce star formation, but a complete understanding of the involved processes is still lacking. The effect of dwarf galaxy mergers is even less clear than their massive counterpart. We present a study on a dwarf merger remnant in the Virgo cluster, VCC322, which might form a triple system with VCC334 and VCC319. We identify a prominent long and straight tail-like substructure that has a size comparable to its host galaxy VCC322. By comparing the color-color ($g-r$ vs. $r-H$) distribution with simple stellar population models, we infer that the metallicity and stellar age of this tail are $Z_\star \sim 0.02~Z_\odot$ and $t_\star \sim 10$~Gyr, respectively. In VCC319, we find a sign of isophotal twisting. This suggests that VCC319 may be subject to tidal interaction. An analysis of the SDSS optical spectra of VCC322 indicates mass- and light-weighted ages of about 10$^{9.8}$ yr and 10$^{7.5}$ yr, respectively, indicating an ongoing star formation activity. However, the star formation in VCC322 seems suppressed when compared to other star-forming dwarfs of comparable stellar masses. Our finding of shock excitation of optical emission lines indicates that interaction-induced shock may contribute to the heating of cold gas and suppression of star formation., Comment: 12 pages, 7 figures. Accepted for publication in ApJ

Recent advances in high-fidelity virtual environments serve as one of the major driving forces for building intelligent embodied agents to perceive, reason and interact with the physical world. Typically, these environments remain unchanged unless agents interact with them. However, in real-world scenarios, agents might also face dynamically changing environments characterized by unexpected events and need to rapidly take action accordingly. To remedy this gap, we propose a new simulated embodied benchmark, called HAZARD, specifically designed to assess the decision-making abilities of embodied agents in dynamic situations. HAZARD consists of three unexpected disaster scenarios, including fire, flood, and wind, and specifically supports the utilization of large language models (LLMs) to assist common sense reasoning and decision-making. This benchmark enables us to evaluate autonomous agents' decision-making capabilities across various pipelines, including reinforcement learning (RL), rule-based, and search-based methods in dynamically changing environments. As a first step toward addressing this challenge using large language models, we further develop an LLM-based agent and perform an in-depth analysis of its promise and challenge of solving these challenging tasks. HAZARD is available at https://vis-www.cs.umass.edu/hazard/., Comment: ICLR 2024. The first two authors contributed equally to this work

Systematic studies have revealed hundreds of ultra-compact dwarf galaxies (UCDs) in the nearby Universe. With half-light radii $r_h$ of approximately 10-100 parsecs and stellar masses $M_*$ $\approx$ $10^6-10^8$ solar masses, UCDs are among the densest known stellar systems. Although similar in appearance to massive globular clusters, the detection of extended stellar envelopes, complex star formation histories, elevated mass-to-light ratio, and supermassive black holes suggest that some UCDs are remnant nuclear star clusters of tidally-stripped dwarf galaxies, or even ancient compact galaxies. However, only a few objects have been found in the transient stage of tidal stripping, and this assumed evolutionary path has never been fully traced by observations. Here we show that 106 galaxies in the Virgo cluster have morphologies that are intermediate between normal, nucleated dwarf galaxies and single-component UCDs, revealing a continuum that fully maps this morphological transition, and fills the `size gap' between star clusters and galaxies. Their spatial distribution and redder color are also consistent with stripped satellite galaxies on their first few pericentric passages around massive galaxies. The `ultra-diffuse' tidal features around several of these galaxies directly show how UCDs are forming through tidal stripping, and that this evolutionary path can include an early phase as a nucleated ultra-diffuse galaxy (UDG). These UCDs represent substantial visible fossil remnants of ancient dwarf galaxies in galaxy clusters, and more low-mass remnants probably remain to be found., Comment: Published in Nature. Accepted on September 15

Supervised Fine-Tuning (SFT) on response demonstrations combined with Reinforcement Learning from Human Feedback (RLHF) constitutes a powerful paradigm for aligning LLM-based AI agents. However, a significant limitation of such an approach is its dependency on high-quality human annotations, making its application to intricate tasks challenging due to difficulties in obtaining consistent response demonstrations and in-distribution response preferences. This paper presents a novel approach, namely SALMON, to align base language models with minimal human supervision, using only a small set of human-defined principles, yet achieving superior performance. Central to our approach is an instructable reward model. Trained on synthetic preference data, this model can generate reward scores based on arbitrary human-defined principles. By merely adjusting these principles during the RL training phase, we gain full control over the preferences with the instructable reward model, subsequently influencing the behavior of the RL-trained policy models, and reducing the reliance on the collection of online human preferences. Applying our method to the LLaMA-2-70b base language model, we developed an AI assistant named Dromedary-2. With only 6 exemplars for in-context learning and 31 human-defined principles, Dromedary-2 significantly surpasses the performance of several state-of-the-art AI systems, including LLaMA-2-Chat-70b, on various benchmark datasets. We have open-sourced the code and model weights to encourage further research into aligning LLM-based AI agents with enhanced supervision efficiency, improved controllability, and scalable oversight., Comment: Previous Title: SALMON: Self-Alignment with Principle-Following Reward Models. Accepted to ICLR 2024. Project page: https://github.com/IBM/SALMON

Green pea galaxies are a special class of star-forming compact galaxies with strong [O III]{\lambda}5007 and considered as analogs of high-redshift Ly{\alpha}-emitting galaxies and potential sources for cosmic reionization. In this paper, we identify 76 strong [O III]{\lambda}5007 compact galaxies at z < 0.35 from DR1613 of the Sloan Digital Sky Survey. These galaxies present relatively low stellar mass, high star formation rate, and low metallicity. Both star-forming main sequence relation (SFMS) and mass-metallicity relation (MZR) are investigated and compared with green pea and blueberry galaxies collected from literature. It is found that our strong [O III] {\lambda}5007 compact galaxies share common properties with those compact galaxies with extreme star formation and show distinct scaling relations in respect to those of normal star-forming galaxies at the same redshift. The slope of SFMS is higher, indicates that strong [O III]{\lambda}5007 compact galaxies might grow faster in stellar mass. The lower MZR implies that they may be less chemically evolved and hence on the early stage of star formation. A further environmental investigation confirms that they inhabit relatively low-density regions. Future largescale spectroscopic surveys will provide more details on their physical origin and evolution., Comment: 12 pages, 8 figures, 1 table. Published in AJ

The physical mechanisms driving starbursts in dwarf galaxies are unclear, and the effects of mergers on star formation in these galaxies are still uncertain. We explore how the merger process affects star formation in metal-poor dwarf galaxies by analyzing high-spatial-resolution ($\sim$ 70 pc) integral field spectrograph observations of ionized gas. We use archival data from the Very Large Telescope/Multi Unit Spectroscopic Explorer to map the spatial distribution of strong emission lines (e.g., $\rm H\beta$, $\rm H\alpha$, $\rm [OIII]\lambda5007$, $\rm [NII]\lambda6583$, etc) in the nearby merging star-forming dwarf galaxy system NGC 4809/4810. We identify approximately 112 star-forming knots scattered among the two galaxies, where the gas-phase metallicity distribution is inhomogeneous and mixing with metal-poor and metal-rich ionized gas. Star-forming knots at the interacting region show lower metallicity, the highest star formation rates (SFRs) and SFR to resolved main-sequence-relation (rMSR) ratios. Ionized gas exhibits an obvious northeast-southwest velocity gradient in NGC 4809, while seemingly mixed in NGC 4810. High virial parameters and the stellar mass-size relation of HII regions indicate that these regions are dominated by direct radiation pressure from massive stars/clusters and persistently expanding. We find two different stellar mass surface density-stellar age relations in NGC 4809 and NGC 4810, and the stellar ages of NGC 4810 are systematically younger than in NGC 4809. Our study suggests that the merging stage of two dwarf galaxies can induce starburst activities at the interaction areas, despite the metal-deficient environment. Considering the high specific SFRs and different stellar ages, we propose that the interaction initially triggered star formation in NGC 4809 and then drove star formation in NGC 4810., Comment: 13 pages, 12 figures; accepted for publication in A&A

In this work, we address challenging multi-agent cooperation problems with decentralized control, raw sensory observations, costly communication, and multi-objective tasks instantiated in various embodied environments. While previous research either presupposes a cost-free communication channel or relies on a centralized controller with shared observations, we harness the commonsense knowledge, reasoning ability, language comprehension, and text generation prowess of LLMs and seamlessly incorporate them into a cognitive-inspired modular framework that integrates with perception, memory, and execution. Thus building a Cooperative Embodied Language Agent CoELA, who can plan, communicate, and cooperate with others to accomplish long-horizon tasks efficiently. Our experiments on C-WAH and TDW-MAT demonstrate that CoELA driven by GPT-4 can surpass strong planning-based methods and exhibit emergent effective communication. Though current Open LMs like LLAMA-2 still underperform, we fine-tune a CoELA with data collected with our agents and show how they can achieve promising performance. We also conducted a user study for human-agent interaction and discovered that CoELA communicating in natural language can earn more trust and cooperate more effectively with humans. Our research underscores the potential of LLMs for future research in multi-agent cooperation. Videos can be found on the project website https://vis-www.cs.umass.edu/Co-LLM-Agents/., Comment: ICLR24. The first two authors contributed equally

The Wide Field Survey Telescope (WFST) is a dedicated photometric surveying facility being built jointly by the University of Science and Technology of China and the Purple Mountain Observatory. It is equipped with a 2.5-meter diameter primary mirror, an active optics system, and a mosaic CCD camera with 0.73 gigapixels on the primary focal plane for high-quality image capture over an FOV of 6.5-square-degree. It is anticipated that WFST will be set up at the Lenghu site in the summer of 2023 and begin to observe the northern sky in four optical bands (u, g, r, and i) with a range of cadences, from hourly/daily in the Deep High-Cadence Survey (DHS) program to semiweekly in the Wide-Field Survey (WFS) program, three months later. During a photometric night, a nominal 30 s exposure in the WFS program will reach a depth of 22.27, 23.32, 22.84, and 22.31 (AB magnitudes) in these four bands, respectively, allowing for the detection of a tremendous amount of transients in the low-z universe and a systematic investigation of the variability of Galactic and extragalactic objects. In the DHS program, intranight 90 s exposures as deep as 23 (u) and 24 mag (g), in combination with target of opportunity follow-ups, will provide a unique opportunity to explore energetic transients in demand for high sensitivities, including the electromagnetic counterparts of gravitational wave events, supernovae within a few hours of their explosions, tidal disruption events and fast, luminous optical transients even beyond a redshift of unity. In addition, the final 6-year co-added images, anticipated to reach g=25.8 mag in WFS or 1.5 mags deeper in DHS, will be of fundamental importance to general Galactic and extragalactic science. The highly uniform legacy surveys of WFST will serve as an indispensable complement to those of LSST that monitor the southern sky., Comment: 48 pages

Recent AI-assistant agents, such as ChatGPT, predominantly rely on supervised fine-tuning (SFT) with human annotations and reinforcement learning from human feedback (RLHF) to align the output of large language models (LLMs) with human intentions, ensuring they are helpful, ethical, and reliable. However, this dependence can significantly constrain the true potential of AI-assistant agents due to the high cost of obtaining human supervision and the related issues on quality, reliability, diversity, self-consistency, and undesirable biases. To address these challenges, we propose a novel approach called SELF-ALIGN, which combines principle-driven reasoning and the generative power of LLMs for the self-alignment of AI agents with minimal human supervision. Our approach encompasses four stages: first, we use an LLM to generate synthetic prompts, and a topic-guided method to augment the prompt diversity; second, we use a small set of human-written principles for AI models to follow, and guide the LLM through in-context learning from demonstrations (of principles application) to produce helpful, ethical, and reliable responses to user's queries; third, we fine-tune the original LLM with the high-quality self-aligned responses so that the resulting model can generate desirable responses for each query directly without the principle set and the demonstrations anymore; and finally, we offer a refinement step to address the issues of overly-brief or indirect responses. Applying SELF-ALIGN to the LLaMA-65b base language model, we develop an AI assistant named Dromedary. With fewer than 300 lines of human annotations (including < 200 seed prompts, 16 generic principles, and 5 exemplars for in-context learning). Dromedary significantly surpasses the performance of several state-of-the-art AI systems, including Text-Davinci-003 and Alpaca, on benchmark datasets with various settings., Comment: Accepted at NeurIPS 2023 (Spotlight). Project page: https://github.com/IBM/Dromedary

The limits of open-ended generative models are unclear, yet increasingly important. What causes them to succeed and what causes them to fail? In this paper, we take a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models. We present a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic. These constraint types are categorized into a set of well-defined constraints that are analyzable by a single prompt. We then systematically create a diverse set of simple, natural, and useful prompts to robustly analyze each individual constraint. Using the GPT-3 text-davinci-002 model as a case study, we generate outputs from our collection of prompts and analyze the model's generative failures. We also show the generalizability of our proposed method on other large models like BLOOM and OPT. Our results and our in-context mitigation strategies reveal open challenges for future research. We have publicly released our code at https://github.com/SALT-NLP/Bound-Cap-LLM., Comment: 27 pages, 13 figures, 11 tables, to be published in EACL 2023 Findings

Persuasion modeling is a key building block for conversational agents. Existing works in this direction are limited to analyzing textual dialogue corpus. We argue that visual signals also play an important role in understanding human persuasive behaviors. In this paper, we introduce the first multimodal dataset for modeling persuasion behaviors. Our dataset includes 199 dialogue transcriptions and videos captured in a multi-player social deduction game setting, 26,647 utterance level annotations of persuasion strategy, and game level annotations of deduction game outcomes. We provide extensive experiments to show how dialogue context and visual signals benefit persuasion strategy prediction. We also explore the generalization ability of language models for persuasion modeling and the role of persuasion strategies in predicting social deduction game outcomes. Our dataset, code, and models can be found at https://persuasion-deductiongame.socialai-data.org., Comment: 17 pages

Generating a Chain of Thought (CoT) has been shown to consistently improve large language model (LLM) performance on a wide range of NLP tasks. However, prior work has mainly focused on logical reasoning tasks (e.g. arithmetic, commonsense QA); it remains unclear whether improvements hold for more diverse types of reasoning, especially in socially situated contexts. Concretely, we perform a controlled evaluation of zero-shot CoT across two socially sensitive domains: harmful questions and stereotype benchmarks. We find that zero-shot CoT reasoning in sensitive domains significantly increases a model's likelihood to produce harmful or undesirable output, with trends holding across different prompt formats and model variants. Furthermore, we show that harmful CoTs increase with model size, but decrease with improved instruction following. Our work suggests that zero-shot CoT should be used with caution on socially important tasks, especially when marginalized groups or sensitive topics are involved., Comment: ACL 2023 Main Conference

Demonstration-based learning has shown great potential in stimulating pretrained language models' ability under limited data scenario. Simply augmenting the input with some demonstrations can significantly improve performance on few-shot NER. However, why such demonstrations are beneficial for the learning process remains unclear since there is no explicit alignment between the demonstrations and the predictions. In this paper, we design pathological demonstrations by gradually removing intuitively useful information from the standard ones to take a deep dive of the robustness of demonstration-based sequence labeling and show that (1) demonstrations composed of random tokens still make the model a better few-shot learner; (2) the length of random demonstrations and the relevance of random tokens are the main factors affecting the performance; (3) demonstrations increase the confidence of model predictions on captured superficial patterns. We have publicly released our code at https://github.com/SALT-NLP/RobustDemo., Comment: 14 pages, EMNLP 2022 Main Conference

The scaling relationship is a fundamental probe of the evolution of galaxies. Using the integral field spectroscopic data from the Mapping Nearby Galaxies at Apache Point Observatory survey, we select 1698 spaxels with significant detection of the auroral emission line \oiii$\lambda$4363 from 52 galaxies to investigate the scaling relationships at the low-metallicity end. We find that our sample's star formation rate is higher and its metallicity is lower in the scaling relationship than the star-forming sequence after removing the contribution of the Fundamental Metallicity Relation.We also find that the stellar ages of our sample are younger ($<$ 1 Gyr) and the stellar metallicities are also lower. Morphological parameters from Deep Learning catalog indicate that our galaxies are more likely to be merger. These results suggest that their low metallicity regions may be related to interaction, the inflow of metal-poor gas may dilute the interstellar medium and form new metal-poor stars in these galaxies during interaction., Comment: 12 pages, 10 figures, accepted in ApJ

We present the first results of an MMT/Hectospec campaign to measure the kinematics of globular clusters (GCs) around M49 -- the brightest galaxy in the Virgo galaxy cluster, which dominates the Virgo B subcluster. The data include kinematic tracers beyond 95 kpc (~5.2 effective radii) for M49 for the first time, enabling us to achieve three key insights reported here. First, beyond ~20'-30' (~100-150 kpc), the GC kinematics sampled along the minor photometric axis of M49 become increasingly hotter, indicating a transition from GCs related to M49 to those representing the Virgo B intra-cluster medium. Second, there is an anomaly in the line-of-sight radial velocity dispersion ($\sigma_{r,los}$) profile in an annulus ~10-15' (~50-90 kpc) from M49 in which the kinematics cool by $\Delta \sigma_{r,los}~150$ km s$^{-1}$ relative to those in- or outward. The kinematic fingerprint of a previous accretion event is hinted at in projected phase-space, and we isolate GCs that both give rise to this feature, and are spatially co-located with two prominent stellar shells in the halo of M49. Third, we find a subsample of GCs with velocities representative of the dwarf galaxy VCC1249 that is currently interacting with M49. The spatial distribution of these GCs closely resembles the morphology of VCC1249's isophotes, indicating that several of these GCs are likely in the act of being stripped from the dwarf during its passage through M49's halo. Taken together, these results point toward the opportunity of witnessing on-going giant halo assembly in the depths of a cluster environment., Comment: 14 pages, 9 figures, accepted for publication in ApJ

Based on the spectroscopic and shear catalogs for SDSS galaxies in the local Universe, we compare optically-selected active galactic nuclei (AGNs) with control star-forming and quiescent galaxies on galactic, inter-halo and larger scales. We find that AGNs are preferentially found in two specific stages of galaxy evolution: star-burst and `green valley' phases, and that the stellar population of their host galaxies is quite independent of stellar mass, different from normal galaxies. Combining galaxy-galaxy lensing and galaxy clustering on large scales, we measure the mass of AGN host halos. The typical halo mass is about $10^{12}h^{-1}\rm M_{\odot}$, similar to the characteristic mass in the stellar mass-halo mass relation (SHMR). For given stellar mass, AGN host galaxies and star-forming galaxies share the same SHMR, while quiescent galaxies have more massive halos. Clustering analysis on halo scales reveals that AGNs are surrounded by a larger number of satellites (with stellar mass down to 1/1000 of the mass of the central galaxy) than star-forming galaxies, and that galaxies with larger stellar velocity dispersion have more satellites. The number of satellites also increase with halo mass, reaching unity around $10^{12}h^{-1}\rm M_{\odot}$. Our results suggest a scenario, in which the interaction of the central galaxy with the satellites triggers an early episode of star burst and AGN activities, followed by multiple AGN cycles driven by the non-axisymmetric structure produced by the interaction. The feedback from the starburst and AGN reduces the amount of cold gas for fueling the central black hole, producing a characteristic halo mass scale, $\sim 10^{12}h^{-1}\rm M_{\odot}$, where the AGN fraction peaks., Comment: accepted by A&A, 17 pages, 11 figures