Unlocking the Future: Exploring Look-Ahead Planning Mechanistic Interpretability in Large Language Models

Planning, as the core module of agents, is crucial in various fields such as embodied agents, web navigation, and tool using. With the development of large language models (LLMs), some researchers treat large language models as intelligent agents to stimulate and evaluate their planning capabilities. However, the planning mechanism is still unclear. In this work, we focus on exploring the look-ahead planning mechanism in large language models from the perspectives of information flow and internal representations. First, we study how planning is done internally by analyzing the multi-layer perception (MLP) and multi-head self-attention (MHSA) components at the last token. We find that the output of MHSA in the middle layers at the last token can directly decode the decision to some extent. Based on this discovery, we further trace the source of MHSA by information flow, and we reveal that MHSA mainly extracts information from spans of the goal states and recent steps. According to information flow, we continue to study what information is encoded within it. Specifically, we explore whether future decisions have been encoded in advance in the representation of flow. We demonstrate that the middle and upper layers encode a few short-term future decisions to some extent when planning is successful. Overall, our research analyzes the look-ahead planning mechanisms of LLMs, facilitating future research on LLMs performing planning tasks.

Beyond Under-Alignment: Atomic Preference Enhanced Factuality Tuning for Large Language Models

Large language models (LLMs) have achieved remarkable success but still tend to generate factually erroneous responses, a phenomenon known as hallucination. A recent trend is to use preference learning to fine-tune models to align with factuality. However, existing work primarily evaluates fine-tuned models on in-domain (ID) datasets and the factuality on out-of-domain (OOD) datasets remains underexplored. In this paper, we conduct a comprehensive evaluation of the factuality of different models tuned by various preference learning algorithms and demonstrate that their performance on OOD datasets either increases minimally or decreases. Subsequently, we reveal that the main cause of model's failure to uphold factuality under a distribution shift is \textbf{under-alignment}, rather than \textbf{over-alignment}, by analyzing the token distribution shift of the models before and after tuning. Finally, we propose \textbf{APEFT} (\textbf{A}tomic \textbf{P}reference \textbf{E}nhanced \textbf{F}actuality \textbf{T}uning), a framework that enhances model's awareness of factuality at the granularity of individual facts. Extensive experiments demonstrate that APEFT improves model performance by an average of $\boldsymbol{3.45\%}$ on both ID and OOD datasets, which is highly effective.

RWKU: Benchmarking Real-World Knowledge Unlearning for Large Language Models

Large language models (LLMs) inevitably memorize sensitive, copyrighted, and harmful knowledge from the training corpus; therefore, it is crucial to erase this knowledge from the models. Machine unlearning is a promising solution for efficiently removing specific knowledge by post hoc modifying models. In this paper, we propose a Real-World Knowledge Unlearning benchmark (RWKU) for LLM unlearning. RWKU is designed based on the following three key factors: (1) For the task setting, we consider a more practical and challenging unlearning setting, where neither the forget corpus nor the retain corpus is accessible. (2) For the knowledge source, we choose 200 real-world famous people as the unlearning targets and show that such popular knowledge is widely present in various LLMs. (3) For the evaluation framework, we design the forget set and the retain set to evaluate the model's capabilities across various real-world applications. Regarding the forget set, we provide four four membership inference attack (MIA) methods and nine kinds of adversarial attack probes to rigorously test unlearning efficacy. Regarding the retain set, we assess locality and utility in terms of neighbor perturbation, general ability, reasoning ability, truthfulness, factuality, and fluency. We conduct extensive experiments across two unlearning scenarios, two models and six baseline methods and obtain some meaningful findings. We release our benchmark and code publicly at http://rwku-bench.github.io for future work.

SimuCourt: Building Judicial Decision-Making Agents with Real-world Judgement Documents

With the development of deep learning, natural language processing technology has effectively improved the efficiency of various aspects of the traditional judicial industry. However, most current efforts focus solely on individual judicial stage, overlooking cross-stage collaboration. As the autonomous agents powered by large language models are becoming increasingly smart and able to make complex decisions in real-world settings, offering new insights for judicial intelligence. In this paper, (1) we introduce SimuCourt, a judicial benchmark that encompasses 420 judgment documents from real-world, spanning the three most common types of judicial cases, and a novel task Judicial Decision-Making to evaluate the judicial analysis and decision-making power of agents. To support this task, we construct a large-scale judicial knowledge base, JudicialKB, with multiple legal knowledge. (2) we propose a novel multi-agent framework, AgentsCourt. Our framework follows the real-world classic court trial process, consisting of court debate simulation, legal information retrieval and judgement refinement to simulate the decision-making of judge. (3) we perform extensive experiments, the results demonstrate that, our framework outperforms the existing advanced methods in various aspects, especially in generating legal grounds, where our model achieves significant improvements of 8.6% and 9.1% F1 score in the first and second instance settings, respectively.

Whispers that Shake Foundations: Analyzing and Mitigating False Premise Hallucinations in Large Language Models

Large Language Models (LLMs) have shown impressive capabilities but still suffer from the issue of hallucinations. A significant type of this issue is the false premise hallucination, which we define as the phenomenon when LLMs generate hallucinated text when confronted with false premise questions. In this paper, we perform a comprehensive analysis of the false premise hallucination and elucidate its internal working mechanism: a small subset of attention heads (which we designate as false premise heads) disturb the knowledge extraction process, leading to the occurrence of false premise hallucination. Based on our analysis, we propose \textbf{FAITH} (\textbf{F}alse premise \textbf{A}ttention head constra\textbf{I}ining for mi\textbf{T}igating \textbf{H}allucinations), a novel and effective method to mitigate false premise hallucinations. It constrains the false premise attention heads during the model inference process. Impressively, extensive experiments demonstrate that constraining only approximately $1\%$ of the attention heads in the model yields a notable increase of nearly $20\%$ of model performance.

Focus on Your Question! Interpreting and Mitigating Toxic CoT Problems in Commonsense Reasoning

Large language models exhibit high-level commonsense reasoning abilities, especially with enhancement methods like Chain-of-Thought (CoT). However, we find these CoT-like methods lead to a considerable number of originally correct answers turning wrong, which we define as the Toxic CoT problem. To interpret and mitigate this problem, we first utilize attribution tracing and causal tracing methods to probe the internal working mechanism of the LLM during CoT reasoning. Through comparisons, we prove that the model exhibits information loss from the question over the shallow attention layers when generating rationales or answers. Based on the probing findings, we design a novel method called RIDERS (Residual decodIng and sERial-position Swap), which compensates for the information deficit in the model from both decoding and serial-position perspectives. Through extensive experiments on multiple commonsense reasoning benchmarks, we validate that this method not only significantly eliminates Toxic CoT problems (decreased by 23.6%), but also effectively improves the model's overall commonsense reasoning performance (increased by 5.5%).

Cutting Off the Head Ends the Conflict: A Mechanism for Interpreting and Mitigating Knowledge Conflicts in Language Models

Recently, retrieval augmentation and tool augmentation have demonstrated a remarkable capability to expand the internal memory boundaries of language models (LMs) by providing external context. However, internal memory and external context inevitably clash, leading to knowledge conflicts within LMs. In this paper, we aim to interpret the mechanism of knowledge conflicts through the lens of information flow, and then mitigate conflicts by precise interventions at the pivotal point. We find there are some attention heads with opposite effects in the later layers, where memory heads can recall knowledge from internal memory, and context heads can retrieve knowledge from external context. Moreover, we reveal that the pivotal point at which knowledge conflicts emerge in LMs is the integration of inconsistent information flows by memory heads and context heads. Inspired by the insights, we propose a novel method called Pruning Head via PatH PatcHing (PH3), which can efficiently mitigate knowledge conflicts by pruning conflicting attention heads without updating model parameters. PH3 can flexibly control eight LMs to use internal memory ($\uparrow$ 44.0%) or external context ($\uparrow$ 38.5%). Moreover, PH3 can also improve the performance of LMs on open-domain QA tasks. We also conduct extensive experiments to demonstrate the cross-model, cross-relation, and cross-format generalization of our method.

Tug-of-War Between Knowledge: Exploring and Resolving Knowledge Conflicts in Retrieval-Augmented Language Models

Retrieval-augmented language models (RALMs) have demonstrated significant potential in refining and expanding their internal memory by retrieving evidence from external sources. However, RALMs will inevitably encounter knowledge conflicts when integrating their internal memory with external sources. Knowledge conflicts can ensnare RALMs in a tug-of-war between knowledge, limiting their practical applicability. In this paper, we focus on exploring and resolving knowledge conflicts in RALMs. First, we present an evaluation framework for assessing knowledge conflicts across various dimensions. Then, we investigate the behavior and preference of RALMs from the following two perspectives: (1) Conflicts between internal memory and external sources: We find that stronger RALMs emerge with the Dunning-Kruger effect, persistently favoring their faulty internal memory even when correct evidence is provided. Besides, RALMs exhibit an availability bias towards common knowledge; (2) Conflicts between truthful, irrelevant and misleading evidence: We reveal that RALMs follow the principle of majority rule, leaning towards placing trust in evidence that appears more frequently. Moreover, we find that RALMs exhibit confirmation bias, and are more willing to choose evidence that is consistent with their internal memory. To solve the challenge of knowledge conflicts, we propose a method called Conflict-Disentangle Contrastive Decoding (CD2) to better calibrate the model's confidence. Experimental results demonstrate that our CD2 can effectively resolve knowledge conflicts in RALMs.