Search

Role-Playing Agents (RPAs) have shown remarkable performance in various applications, yet they often struggle to recognize and appropriately respond to hard queries that conflict with their role-play knowledge. To investigate RPAs' performance when faced with different types of conflicting requests, we develop an evaluation benchmark that includes contextual knowledge conflicting requests, parametric knowledge conflicting requests, and non-conflicting requests to assess RPAs' ability to identify conflicts and refuse to answer appropriately without over-refusing. Through extensive evaluation, we find that most RPAs behave significant performance gaps toward different conflict requests. To elucidate the reasons, we conduct an in-depth representation-level analysis of RPAs under various conflict scenarios. Our findings reveal the existence of rejection regions and direct response regions within the model's forwarding representation, and thus influence the RPA's final response behavior. Therefore, we introduce a lightweight representation editing approach that conveniently shifts conflicting requests to the rejection region, thereby enhancing the model's refusal accuracy. The experimental results validate the effectiveness of our editing method, improving RPAs' refusal ability of conflicting requests while maintaining their general role-playing capabilities.

Backpropagation is a cornerstone algorithm in training neural networks for supervised learning, which uses a gradient descent method to update network weights by minimizing the discrepancy between actual and desired outputs. Despite its pivotal role in propelling deep learning advancements, the biological plausibility of backpropagation is questioned due to its requirements for weight symmetry, global error computation, and dual-phase training. To address this long-standing challenge, many studies have endeavored to devise biologically plausible training algorithms. However, a fully biologically plausible algorithm for training multilayer neural networks remains elusive, and interpretations of biological plausibility vary among researchers. In this study, we establish criteria for biological plausibility that a desirable learning algorithm should meet. Using these criteria, we evaluate a range of existing algorithms considered to be biologically plausible, including Hebbian learning, spike-timing-dependent plasticity, feedback alignment, target propagation, predictive coding, forward-forward algorithm, perturbation learning, local losses, and energy-based learning. Additionally, we empirically evaluate these algorithms across diverse network architectures and datasets. We compare the feature representations learned by these algorithms with brain activity recorded by non-invasive devices under identical stimuli, aiming to identify which algorithm can most accurately replicate brain activity patterns. We are hopeful that this study could inspire the development of new biologically plausible algorithms for training multilayer networks, thereby fostering progress in both the fields of neuroscience and machine learning.

Conventional federated learning primarily aims to secure the privacy of data distributed across multiple edge devices, with the global model dispatched to edge devices for parameter updates during the learning process. However, the development of large language models (LLMs) requires substantial data and computational resources, rendering them valuable intellectual properties for their developers and owners. To establish a mechanism that protects both data and model privacy in a federated learning context, we introduce a method that just needs to distribute a quantized version of the model's parameters during training. This method enables accurate gradient estimations for parameter updates while preventing clients from accessing a model whose performance is comparable to the centrally hosted one. Moreover, we combine this quantization strategy with LoRA, a popular and parameter-efficient fine-tuning method, to significantly reduce communication costs in federated learning. The proposed framework, named \textsc{FedLPP}, successfully ensures both data and model privacy in the federated learning context. Additionally, the learned central model exhibits good generalization and can be trained in a resource-efficient manner.

The success of Reinforcement Learning from Human Feedback (RLHF) in language model alignment is critically dependent on the capability of the reward model (RM). However, as the training process progresses, the output distribution of the policy model shifts, leading to the RM's reduced ability to distinguish between responses. This issue is further compounded when the RM, trained on a specific data distribution, struggles to generalize to examples outside of that distribution. These two issues can be united as a challenge posed by the shifted distribution of the environment. To surmount this challenge, we introduce MetaRM, a method leveraging meta-learning to align the RM with the shifted environment distribution. MetaRM is designed to train the RM by minimizing data loss, particularly for data that can improve the differentiation ability to examples of the shifted target distribution. Extensive experiments demonstrate that MetaRM significantly improves the RM's distinguishing ability in iterative RLHF optimization, and also provides the capacity to identify subtle differences in out-of-distribution samples., Comment: 11 pages, 6 figures. arXiv admin note: text overlap with arXiv:2401.06080

Parameter Efficient Fine-Tuning (PEFT) techniques have drawn significant attention due to their ability to yield competitive results while updating only a small portion of the adjustable parameters. However, existing PEFT methods pose challenges in hyperparameter selection, such as choosing the rank for LoRA or Adapter, or specifying the length of soft prompts. To address these challenges, we propose a novel fine-tuning approach for neural models, named Representation EDiting (RED), which modifies the representations generated at some layers through the application of scaling and biasing operations. While existing PEFT methods still demonstrate over-parameterization that could potentially undermine the generalization ability acquired from pre-training, RED can substantially reduce the number of trainable parameters by a factor of 25, 700 compared to full parameter fine-tuning and by a factor of 32 relative to LoRA. Remarkably, RED achieves results comparable or superior to both full parameter fine-tuning and other PEFT methods. Extensive experiments across various model architectures and scales, including RoBERTa, GPT-2, T5, and LLaMA-2, have demonstrated the effectiveness and efficiency of RED1, thereby positioning it as a promising PEFT strategy for large-scale neural models.

The recent surge in jailbreaking attacks has revealed significant vulnerabilities in Large Language Models (LLMs) when exposed to malicious inputs. While various defense strategies have been proposed to mitigate these threats, there has been limited research into the underlying mechanisms that make LLMs vulnerable to such attacks. In this study, we suggest that the self-safeguarding capability of LLMs is linked to specific activity patterns within their representation space. Although these patterns have little impact on the semantic content of the generated text, they play a crucial role in shaping LLM behavior under jailbreaking attacks. Our findings demonstrate that these patterns can be detected with just a few pairs of contrastive queries. Extensive experimentation shows that the robustness of LLMs against jailbreaking can be manipulated by weakening or strengthening these patterns. Further visual analysis provides additional evidence for our conclusions, providing new insights into the jailbreaking phenomenon. These findings highlight the importance of addressing the potential misuse of open-source LLMs within the community., Comment: Accepted by COLING 2025

Aligning large language models (LLMs) with human preferences is crucial for enhancing their utility in terms of helpfulness, truthfulness, safety, harmlessness, and interestingness. Existing methods for achieving this alignment often involves employing reinforcement learning from human feedback (RLHF) to fine-tune LLMs based on human labels assessing the relative quality of model responses. Nevertheless, RLHF is susceptible to instability during fine-tuning and presents challenges in implementation.Drawing inspiration from the emerging field of representation engineering (RepE), this study aims to identify relevant representations for high-level human preferences embedded in patterns of activity within an LLM, and achieve precise control of model behavior by transforming its representations. This novel approach, denoted as Representation Alignment from Human Feedback (RAHF), proves to be effective, computationally efficient, and easy to implement.Extensive experiments demonstrate the efficacy of RAHF in not only capturing but also manipulating representations to align with a broad spectrum of human preferences or values, rather than being confined to a singular concept or function (e.g. honesty or bias). RAHF's versatility in accommodating diverse human preferences shows its potential for advancing LLM performance.

Personality plays a pivotal role in shaping human expression patterns, thus regulating the personality of large language models (LLMs) holds significant potential in enhancing the user experience of LLMs. Previous methods either relied on fine-tuning LLMs on specific corpora or necessitated manually crafted prompts to elicit specific personalities from LLMs. However, the former approach is inefficient and costly, while the latter cannot precisely manipulate personality traits at a fine-grained level. To address the above challenges, we have employed a novel Unsupervisedly-Built Personalized Lexicons (UBPL) in a pluggable manner during the decoding phase of LLMs to manipulate their personality traits. UBPL is a lexicon built through an unsupervised approach from a situational judgment test dataset (SJTs4LLM). Users can utilize UBPL to adjust the probability vectors of predicted words in the decoding phase of LLMs, thus influencing the personality expression of LLMs. Extensive experimentation demonstrates the remarkable effectiveness and pluggability of our method for fine-grained manipulation of LLM's personality., Comment: Work in progress

Spiking Neural Networks (SNNs) have emerged as a promising alternative to conventional Artificial Neural Networks (ANNs), demonstrating comparable performance in both visual and linguistic tasks while offering the advantage of improved energy efficiency. Despite these advancements, the integration of linguistic and visual features into a unified representation through spike trains poses a significant challenge, and the application of SNNs to multimodal scenarios remains largely unexplored. This paper presents SpikeCLIP, a novel framework designed to bridge the modality gap in spike-based computation. Our approach employs a two-step recipe: an ``alignment pre-training'' to align features across modalities, followed by a ``dual-loss fine-tuning'' to refine the model's performance. Extensive experiments reveal that SNNs achieve results on par with ANNs while substantially reducing energy consumption across various datasets commonly used for multimodal model evaluation. Furthermore, SpikeCLIP maintains robust image classification capabilities, even when dealing with classes that fall outside predefined categories. This study marks a significant advancement in the development of energy-efficient and biologically plausible multimodal learning systems.

Spiking neural networks (SNNs) offer a promising avenue to implement deep neural networks in a more energy-efficient way. However, the network architectures of existing SNNs for language tasks are still simplistic and relatively shallow, and deep architectures have not been fully explored, resulting in a significant performance gap compared to mainstream transformer-based networks such as BERT. To this end, we improve a recently-proposed spiking Transformer (i.e., Spikformer) to make it possible to process language tasks and propose a two-stage knowledge distillation method for training it, which combines pre-training by distilling knowledge from BERT with a large collection of unlabelled texts and fine-tuning with task-specific instances via knowledge distillation again from the BERT fine-tuned on the same training examples. Through extensive experimentation, we show that the models trained with our method, named SpikeBERT, outperform state-of-the-art SNNs and even achieve comparable results to BERTs on text classification tasks for both English and Chinese with much less energy consumption. Our code is available at https://github.com/Lvchangze/SpikeBERT.