Related papers: Do Large Language Models Get Caught in Hofstadter-…
Large Language Models are increasingly proposed as cognitive components for robotic systems, yet their opaque decision processes make it difficult to explain success or failure in closed-loop embodied tasks. Following an empirical AI…
Explanations of the replication crisis often emphasize misconduct, questionable research practices, or incentive misalignment, implying that behavioral reform is sufficient. This paper argues that a substantial component is architectural:…
Large language models (LLMs) are being increasingly adopted in the software engineering domain, yet the robustness of their grasp on core software design concepts remains unclear. We conduct an empirical study to systematically evaluate…
Causal discovery is a cornerstone of scientific reasoning, yet whether large language models can perform it reliably remains an open question. Recent benchmarks show that even fine-tuned models plateau on simple causal graphs and degrade as…
Large language models trained under diverse objectives and architectures have been shown to develop increasingly similar internal representations, an observation formalized as the Platonic Representation Hypothesis. Whether this…
Large Vision-Language Models (LVLMs) have shown remarkable performance on many visual-language tasks. However, these models still suffer from multimodal hallucination, which means the generation of objects or content that violates the…
Large Language Models (LLMs) have demonstrated great capabilities in natural language understanding and generation, largely attributed to the intricate alignment process using human feedback. While alignment has become an essential training…
Safety alignment in large language models is typically evaluated under isolated queries, yet real-world use is inherently multi-turn. Although multi-turn jailbreaks are empirically effective, the structure of conversational safety failure…
Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that…
Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. However, their tendency to exhibit sycophantic behavior - excessively agreeing with or flattering users - poses…
Large Language Models (LLMs) have revolutionized artificial intelligence, demonstrating remarkable computational power and linguistic capabilities. However, these models are inherently prone to various biases stemming from their training…
Insensitivity to semantically-preserving variations of prompts (paraphrases) is crucial for reliable behavior and real-world deployment of large language models. However, language models exhibit significant performance degradation when…
This paper proposes the External Hippocampus framework, which models language model reasoning from a cognitive dynamics perspective as the flow of information energy in semantic space. Unlike traditional weight-space optimization methods,…
Evaluating the capability of Large Language Models (LLMs) in following instructions has heavily relied on a powerful LLM as the judge, introducing unresolved biases that deviate the judgments from human judges. In this work, we reevaluate…
Large Language Models (LLMs) are trained on large corpora written by humans and demonstrate high performance on various tasks. However, as humans are susceptible to cognitive biases, which can result in irrational judgments, LLMs can also…
Large Language Models (LLMs) are expected to provide helpful and harmless responses, yet they often exhibit sycophancy--conforming to user beliefs regardless of factual accuracy or ethical soundness. Prior research on sycophancy has…
Standard accuracy on binary reasoning benchmarks hides critical failure modes: prior collapse, inconsistency under paraphrase, and inability to reason about parameter-dependent dynamics. We present ChaosBench-Logic v2, a 40,886-question…
Large language models (LLMs) now write code in settings where misreading a single word can break safety or cost money, yet we still expect them to overlook stray typos. To probe where useful robustness ends and harmful insensitivity begins,…
Large language models (LLMs) are increasingly explored for their reasoning capabilities, yet their ability to perform structured, constraint-based optimization from natural language remains insufficiently understood. This study evaluates…
Intrinsic self-correction in Large Language Models (LLMs) frequently fails in open-ended reasoning tasks due to ``hallucination snowballing,'' a phenomenon in which models recursively justify early errors during free-text reflection. While…