In zero-shot image recognition tasks, humans demonstrate remarkable flexibility in classifying unseen categories by composing known simpler concepts. However, existing vision-language models (VLMs), despite achieving significant progress through large-scale natural language supervision, often underperform in real-world applications because of sub-optimal prompt engineering and the inability to adapt effectively to target classes. To address these issues, we propose a Concept-guided Human-like Bayesian Reasoning (CHBR) framework. Grounded in Bayes' theorem, CHBR models the concept used in human image recognition as latent variables and formulates this task by summing across potential concepts, weighted by a prior distribution and a likelihood function. To tackle the intractable computation over an infinite concept space, we introduce an importance sampling algorithm that iteratively prompts large language models (LLMs) to generate discriminative concepts, emphasizing inter-class differences. We further propose three heuristic approaches involving Average Likelihood, Confidence Likelihood, and Test Time Augmentation (TTA) Likelihood, which dynamically refine the combination of concepts based on the test image. Extensive evaluations across fifteen datasets demonstrate that CHBR consistently outperforms existing state-of-the-art zero-shot generalization methods.

Retrieval-Augmented Generation (RAG) is a popular approach for enhancing Large Language Models (LLMs) by addressing their limitations in verifying facts and answering knowledge-intensive questions. As the research in LLM extends their capability to handle input modality other than text, e.g. image, several multimodal RAG benchmarks are proposed. Nonetheless, they mainly use textual knowledge bases as the primary source of evidences for augmentation. There still lack benchmarks designed to evaluate images as augmentation in RAG systems and how they leverage visual knowledge. We propose Visual-RAG, a novel Question Answering benchmark that emphasizes visual knowledge intensive questions. Unlike prior works relying on text-based evidence, Visual-RAG necessitates text-to-image retrieval and integration of relevant clue images to extract visual knowledge as evidence. With Visual-RAG, we evaluate 5 open-sourced and 3 proprietary Multimodal LLMs (MLLMs), revealing that images can serve as good evidence in RAG; however, even the SoTA models struggle with effectively extracting and utilizing visual knowledge

Integrating large language models (LLMs) with rule-based reasoning offers a powerful solution for improving the flexibility and reliability of Knowledge Base Completion (KBC). Traditional rule-based KBC methods offer verifiable reasoning yet lack flexibility, while LLMs provide strong semantic understanding yet suffer from hallucinations. With the aim of combining LLMs' understanding capability with the logical and rigor of rule-based approaches, we propose a novel framework consisting of a Subgraph Extractor, an LLM Proposer, and a Rule Reasoner. The Subgraph Extractor first samples subgraphs from the KB. Then, the LLM uses these subgraphs to propose diverse and meaningful rules that are helpful for inferring missing facts. To effectively avoid hallucination in LLMs' generations, these proposed rules are further refined by a Rule Reasoner to pinpoint the most significant rules in the KB for Knowledge Base Completion. Our approach offers several key benefits: the utilization of LLMs to enhance the richness and diversity of the proposed rules and the integration with rule-based reasoning to improve reliability. Our method also demonstrates strong performance across diverse KB datasets, highlighting the robustness and generalizability of the proposed framework.

Circuit discovery with edge-level ablation has become a foundational framework for mechanism interpretability of language models. However, its focus on individual edges often overlooks the sequential, path-level causal relationships that underpin complex behaviors, thus potentially leading to misleading or incomplete circuit discoveries. To address this issue, we propose a novel path-level circuit discovery framework capturing how behaviors emerge through interconnected linear chain and build towards complex behaviors. Our framework is constructed upon a fully-disentangled linear combinations of ``memory circuits'' decomposed from the original model. To discover functional circuit paths, we leverage a 2-step pruning strategy by first reducing the computational graph to a faithful and minimal subgraph and then applying causal mediation to identify common paths of a specific skill, termed as skill paths. In contrast to circuit graph from existing works, we focus on the complete paths of a generic skill rather than on the fine-grained responses to individual components of the input. To demonstrate this, we explore three generic language skills, namely Previous Token Skill, Induction Skill and In-Context Learning Skill using our framework and provide more compelling evidence to substantiate stratification and inclusiveness of these skills.

Fact-checking pipelines increasingly adopt the Decompose-Then-Verify paradigm, where texts are broken down into smaller claims for individual verification and subsequently combined for a veracity decision. While decomposition is widely-adopted in such pipelines, its effects on final fact-checking performance remain underexplored. Some studies have reported improvements from decompostition, while others have observed performance declines, indicating its inconsistent impact. To date, no comprehensive analysis has been conducted to understand this variability. To address this gap, we present an in-depth analysis that explicitly examines the impact of decomposition on downstream verification performance. Through error case inspection and experiments, we introduce a categorization of decomposition errors and reveal a trade-off between accuracy gains and the noise introduced through decomposition. Our analysis provides new insights into understanding current system's instability and offers guidance for future studies toward improving claim decomposition in fact-checking pipelines.

Large Language Models (LLMs) have demonstrated impressive in-context learning (ICL) capabilities from few-shot demonstration exemplars. While recent learning-based demonstration selection methods have proven beneficial to ICL by choosing more useful exemplars, their underlying mechanisms are opaque, hindering efforts to address limitations such as high training costs and poor generalization across tasks. These methods generally assume the selection process captures similarities between the exemplar and the target instance, however, it remains unknown what kinds of similarities are captured and vital to performing ICL. To dive into this question, we analyze the working mechanisms of the learning-based demonstration selection methods and empirically identify two important factors related to similarity measurement: 1) The ability to integrate different levels of task-agnostic text similarities between the input of exemplars and test cases enhances generalization power across different tasks. 2) Incorporating task-specific labels when measuring the similarities significantly improves the performance on each specific task. We validate these two findings through extensive quantitative and qualitative analyses across ten datasets and various LLMs. Based on our findings, we introduce two effective yet simplified exemplar selection methods catering to task-agnostic and task-specific demands, eliminating the costly LLM inference overhead.

Emergence, broadly conceptualized as the ``intelligent'' behaviors of LLMs, has recently been studied and proved challenging to quantify due to the lack of a measurable definition. Most commonly, it has been estimated statistically through model performances across extensive datasets and tasks, which consumes significant resources. In addition, such estimation is difficult to interpret and may not accurately reflect the models' intrinsic emergence. In this work, we propose a quantifiable solution for estimating emergence. Inspired by emergentism in dynamics, we quantify the strength of emergence by comparing the entropy reduction of the macroscopic (semantic) level with that of the microscopic (token) level, both of which are derived from the representations within the transformer block. Using a low-cost estimator, our quantification method demonstrates consistent behaviors across a suite of LMs (GPT-2, GEMMA, etc.) under both in-context learning (ICL) and natural sentences. Empirical results show that (1) our method gives consistent measurements which align with existing observations based on performance metrics, validating the effectiveness of our emergence quantification; (2) our proposed metric uncovers novel emergence patterns such as the correlations between the variance of our metric and the number of ``shots'' in ICL, which further suggests a new way of interpreting hallucinations in LLMs; (3) we offer a potential solution towards estimating the emergence of larger and closed-resource LMs via smaller LMs like GPT-2. Our codes are available at: https://github.com/Zodiark-ch/Emergence-of-LLMs/.

Large Language Models (LLMs) are emerging as promising approaches to enhance session-based recommendation (SBR), where both prompt-based and fine-tuning-based methods have been widely investigated to align LLMs with SBR. However, the former methods struggle with optimal prompts to elicit the correct reasoning of LLMs due to the lack of task-specific feedback, leading to unsatisfactory recommendations. Although the latter methods attempt to fine-tune LLMs with domain-specific knowledge, they face limitations such as high computational costs and reliance on open-source backbones. To address such issues, we propose a Reflective Reinforcement Large Language Model (Re2LLM) for SBR, guiding LLMs to focus on specialized knowledge essential for more accurate recommendations effectively and efficiently. In particular, we first design the Reflective Exploration Module to effectively extract knowledge that is readily understandable and digestible by LLMs. To be specific, we direct LLMs to examine recommendation errors through self-reflection and construct a knowledge base (KB) comprising hints capable of rectifying these errors. To efficiently elicit the correct reasoning of LLMs, we further devise the Reinforcement Utilization Module to train a lightweight retrieval agent. It learns to select hints from the constructed KB based on the task-specific feedback, where the hints can serve as guidance to help correct LLMs reasoning for better recommendations. Extensive experiments on multiple real-world datasets demonstrate that our method consistently outperforms state-of-the-art methods.

In-context Learning (ICL) has emerged as a powerful capability alongside the development of scaled-up large language models (LLMs). By instructing LLMs using few-shot demonstrative examples, ICL enables them to perform a wide range of tasks without updating millions of parameters. However, the precise contributions of demonstrations towards improving end-task performance have not been thoroughly investigated in recent analytical studies. In this paper, we empirically decompose the overall performance of ICL into three dimensions, label space, format, and discrimination, and we evaluate four general-purpose LLMs across a diverse range of tasks. Counter-intuitively, we find that the demonstrations have a marginal impact on provoking discriminative knowledge of language models. However, ICL exhibits significant efficacy in regulating the label space and format which helps LLMs to respond in desired label words. We then demonstrate this ability functions similar to detailed instructions for LLMs to follow. We additionally provide an in-depth analysis of the mechanism of retrieval helping with ICL and find that retrieving the most semantically similar examples notably boosts model's discriminative capability.

Dense retrievers and retrieval-augmented language models have been widely used in various NLP applications. Despite being designed to deliver reliable and secure outcomes, the vulnerability of retrievers to potential attacks remains unclear, raising concerns about their security. In this paper, we introduce a novel scenario where the attackers aim to covertly disseminate targeted misinformation, such as hate speech or advertisement, through a retrieval system. To achieve this, we propose a perilous backdoor attack triggered by grammar errors in dense passage retrieval. Our approach ensures that attacked models can function normally for standard queries but are manipulated to return passages specified by the attacker when users unintentionally make grammatical mistakes in their queries. Extensive experiments demonstrate the effectiveness and stealthiness of our proposed attack method. When a user query is error-free, our model consistently retrieves accurate information while effectively filtering out misinformation from the top-k results. However, when a query contains grammar errors, our system shows a significantly higher success rate in fetching the targeted content.