Dissecting Misalignment of Multimodal Large Language Models via Influence Function

Multi-modal Large Language models (MLLMs) are always trained on data from diverse and unreliable sources, which may contain misaligned or mislabeled text-image pairs. This frequently causes robustness issues and hallucinations, leading to performance degradation. Data valuation is an efficient way to detect and trace these misalignments. Nevertheless, existing methods are computationally expensive for MLLMs. While computationally efficient, the classical influence functions are inadequate for contrastive learning models because they were originally designed for pointwise loss. Additionally, contrastive learning involves minimizing the distance between the modalities of positive samples and maximizing the distance between the modalities of negative samples. This requires us to evaluate the influence of samples from both perspectives. To tackle these challenges, we introduce the Extended Influence Function for Contrastive Loss (ECIF), an influence function crafted for contrastive loss. ECIF considers both positive and negative samples and provides a closed-form approximation of contrastive learning models, eliminating the need for retraining. Building upon ECIF, we develop a series of algorithms for data evaluation in MLLM, misalignment detection, and misprediction trace-back tasks. Experimental results demonstrate our ECIF advances the transparency and interpretability of MLLMs by offering a more accurate assessment of data impact and model alignment compared to traditional baseline methods.

SMoA: Improving Multi-agent Large Language Models with Sparse Mixture-of-Agents

While multi-agent systems have been shown to significantly enhance the performance of Large Language Models (LLMs) across various tasks and applications, the dense interaction between scaling agents potentially hampers their efficiency and diversity. To address these challenges, we draw inspiration from the sparse mixture-of-agents (SMoE) and propose a sparse mixture-of-agents (SMoA) framework to improve the efficiency and diversity of multi-agent LLMs. Unlike completely connected structures, SMoA introduces novel Response Selection and Early Stopping mechanisms to sparsify information flows among individual LLM agents, striking a balance between performance and efficiency. Additionally, inspired by the expert diversity principle in SMoE frameworks for workload balance between experts, we assign distinct role descriptions to each LLM agent, fostering diverse and divergent thinking. Extensive experiments on reasoning, alignment, and fairness benchmarks demonstrate that SMoA achieves performance comparable to traditional mixture-of-agents approaches but with significantly lower computational costs. Further analysis reveals that SMoA is more stable, has a greater capacity to scale, and offers considerable potential through hyper-parameter optimization. Code and data will be available at: https://github.com/David-Li0406/SMoA.

Towards Multi-dimensional Explanation Alignment for Medical Classification

The lack of interpretability in the field of medical image analysis has significant ethical and legal implications. Existing interpretable methods in this domain encounter several challenges, including dependency on specific models, difficulties in understanding and visualization, as well as issues related to efficiency. To address these limitations, we propose a novel framework called Med-MICN (Medical Multi-dimensional Interpretable Concept Network). Med-MICN provides interpretability alignment for various angles, including neural symbolic reasoning, concept semantics, and saliency maps, which are superior to current interpretable methods. Its advantages include high prediction accuracy, interpretability across multiple dimensions, and automation through an end-to-end concept labeling process that reduces the need for extensive human training effort when working with new datasets. To demonstrate the effectiveness and interpretability of Med-MICN, we apply it to four benchmark datasets and compare it with baselines. The results clearly demonstrate the superior performance and interpretability of our Med-MICN.

Private Language Models via Truncated Laplacian Mechanism

Deep learning models for NLP tasks are prone to variants of privacy attacks. To prevent privacy leakage, researchers have investigated word-level perturbations, relying on the formal guarantees of differential privacy (DP) in the embedding space. However, many existing approaches either achieve unsatisfactory performance in the high privacy regime when using the Laplacian or Gaussian mechanism, or resort to weaker relaxations of DP that are inferior to the canonical DP in terms of privacy strength. This raises the question of whether a new method for private word embedding can be designed to overcome these limitations. In this paper, we propose a novel private embedding method called the high dimensional truncated Laplacian mechanism. Specifically, we introduce a non-trivial extension of the truncated Laplacian mechanism, which was previously only investigated in one-dimensional space cases. Theoretically, we show that our method has a lower variance compared to the previous private word embedding methods. To further validate its effectiveness, we conduct comprehensive experiments on private embedding and downstream tasks using three datasets. Remarkably, even in the high privacy regime, our approach only incurs a slight decrease in utility compared to the non-private scenario.

Dissecting Fine-Tuning Unlearning in Large Language Models

Fine-tuning-based unlearning methods prevail for preventing targeted harmful, sensitive, or copyrighted information within large language models while preserving overall capabilities. However, the true effectiveness of these methods is unclear. In this paper, we delve into the limitations of fine-tuning-based unlearning through activation patching and parameter restoration experiments. Our findings reveal that these methods alter the model's knowledge retrieval process, rather than genuinely erasing the problematic knowledge embedded in the model parameters. Furthermore, behavioral tests demonstrate that the unlearning mechanisms inevitably impact the global behavior of the models, affecting unrelated knowledge or capabilities. Our work advocates the development of more resilient unlearning techniques for truly erasing knowledge. Our code is released at https://github.com/yihuaihong/Dissecting-FT-Unlearning.

Faithful Interpretation for Graph Neural Networks

Currently, attention mechanisms have garnered increasing attention in Graph Neural Networks (GNNs), such as Graph Attention Networks (GATs) and Graph Transformers (GTs). It is not only due to the commendable boost in performance they offer but also its capacity to provide a more lucid rationale for model behaviors, which are often viewed as inscrutable. However, Attention-based GNNs have demonstrated instability in interpretability when subjected to various sources of perturbations during both training and testing phases, including factors like additional edges or nodes. In this paper, we propose a solution to this problem by introducing a novel notion called Faithful Graph Attention-based Interpretation (FGAI). In particular, FGAI has four crucial properties regarding stability and sensitivity to interpretation and final output distribution. Built upon this notion, we propose an efficient methodology for obtaining FGAI, which can be viewed as an ad hoc modification to the canonical Attention-based GNNs. To validate our proposed solution, we introduce two novel metrics tailored for graph interpretation assessment. Experimental results demonstrate that FGAI exhibits superior stability and preserves the interpretability of attention under various forms of perturbations and randomness, which makes FGAI a more faithful and reliable explanation tool.

Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing

The locate-then-edit paradigm has shown significant promise for knowledge editing (KE) in Large Language Models (LLMs). While previous methods perform well on single-hop fact recall tasks, they consistently struggle with multi-hop factual recall tasks involving newly edited knowledge. In this paper, leveraging tools in mechanistic interpretability, we first identify that in multi-hop tasks, LLMs tend to retrieve implicit subject knowledge from deeper MLP layers, unlike single-hop tasks, which rely on earlier layers. This distinction explains the poor performance of current methods in multi-hop queries, as they primarily focus on editing shallow layers, leaving deeper layers unchanged. To address this, we propose IFMET, a novel locate-then-edit KE approach designed to edit both shallow and deep MLP layers. IFMET employs multi-hop editing prompts and supplementary sets to locate and modify knowledge across different reasoning stages. Experimental results demonstrate that IFMET significantly improves performance on multi-hop factual recall tasks, effectively overcoming the limitations of previous locate-then-edit methods.

CAT: Concept-level backdoor ATtacks for Concept Bottleneck Models

Despite the transformative impact of deep learning across multiple domains, the inherent opacity of these models has driven the development of Explainable Artificial Intelligence (XAI). Among these efforts, Concept Bottleneck Models (CBMs) have emerged as a key approach to improve interpretability by leveraging high-level semantic information. However, CBMs, like other machine learning models, are susceptible to security threats, particularly backdoor attacks, which can covertly manipulate model behaviors. Understanding that the community has not yet studied the concept level backdoor attack of CBM, because of "Better the devil you know than the devil you don't know.", we introduce CAT (Concept-level Backdoor ATtacks), a methodology that leverages the conceptual representations within CBMs to embed triggers during training, enabling controlled manipulation of model predictions at inference time. An enhanced attack pattern, CAT+, incorporates a correlation function to systematically select the most effective and stealthy concept triggers, thereby optimizing the attack's impact. Our comprehensive evaluation framework assesses both the attack success rate and stealthiness, demonstrating that CAT and CAT+ maintain high performance on clean data while achieving significant targeted effects on backdoored datasets. This work underscores the potential security risks associated with CBMs and provides a robust testing methodology for future security assessments.

What makes your model a low-empathy or warmth person: Exploring the Origins of Personality in LLMs

Large language models (LLMs) have demonstrated remarkable capabilities in generating human-like text and exhibiting personality traits similar to those in humans. However, the mechanisms by which LLMs encode and express traits such as agreeableness and impulsiveness remain poorly understood. Drawing on the theory of social determinism, we investigate how long-term background factors, such as family environment and cultural norms, interact with short-term pressures like external instructions, shaping and influencing LLMs' personality traits. By steering the output of LLMs through the utilization of interpretable features within the model, we explore how these background and pressure factors lead to changes in the model's traits without the need for further fine-tuning. Additionally, we suggest the potential impact of these factors on model safety from the perspective of personality.

Understanding Reasoning in Chain-of-Thought from the Hopfieldian View

Large Language Models have demonstrated remarkable abilities across various tasks, with Chain-of-Thought (CoT) prompting emerging as a key technique to enhance reasoning capabilities. However, existing research primarily focuses on improving performance, lacking a comprehensive framework to explain and understand the fundamental factors behind CoT's success. To bridge this gap, we introduce a novel perspective grounded in the Hopfieldian view of cognition in cognitive neuroscience. We establish a connection between CoT reasoning and key cognitive elements such as stimuli, actions, neural populations, and representation spaces. From our view, we can understand the reasoning process as the movement between these representation spaces. Building on this insight, we develop a method for localizing reasoning errors in the response of CoTs. Moreover, we propose the Representation-of-Thought (RoT) framework, which leverages the robustness of low-dimensional representation spaces to enhance the robustness of the reasoning process in CoTs. Experimental results demonstrate that RoT improves the robustness and interpretability of CoT reasoning while offering fine-grained control over the reasoning process.