The paper "Illusory Pattern Perception Drives Spurious Inference in Large Language Models" reports that large language models (LLMs) exhibit a cognitive bias similar to humans’ tendency to see patterns where none exist. By adapting classic psychological experiments to three tasks, the authors find that LLMs often over‑associate frequent positive attributes with majority groups or large organizations and are prone to constructing causal narratives from ambiguous events. Using a Sparse Autoencoder‑based interpretability framework, the study links these illusory perceptions to holistic frequency perception and analytic cognitive orientation within the models.
arXiv:2601.08058v2 Announce Type: replace-cross Abstract: Chain-of-Thought (CoT) prompting often improves the reasoning performance of large language models (LLMs), but the internal signal that trigg...
arXiv:2508.21285v4 Announce Type: replace-cross Abstract: We use sparse autoencoders (SAEs) to "brain scan" large language models, mapping internal activations to interpretable concepts and enabling...
arXiv:2503. 08884v3 Announce Type: replace-cross Abstract: Unimodal vision models are known to rely on spurious correlations, but it remains unclear to what extent Multimodal Large Language Models (MLLMs) exhibit similar biases despite language supervision.
arXiv:2603. 01437v2 Announce Type: replace Abstract: As chain of thought (CoT) has become central to scaling reasoning capabilities in large language models (LLMs), it has also emerged as a promising tool for interpretability, suggesting the opportunity to understand model decisions through verbalized reasoning.
arXiv:2608.22232v1 Announce Type: new Abstract: Real-world situation appearances can deviate from their underlying physical states, challenging the reliability of multimodal large language models (ML...
arXiv:2609.38802v1 Announce Type: cross Abstract: Uncontrolled repetition can prolong autoregressive generation in large language models (LLMs) and enable resource consumption attacks. Prior analyses...
arXiv:2608. 10537v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) have helped uncover mechanistic explanations for LLM behaviours such as reasoning, jailbreaking etc.
The paper compares latent representations in Selective State Space Models (SSMs) like Mamba and Transformers such as Pythia using Sparse Autoencoders. Across a 10‑million token corpus, 99.98% of Mamba features align closely with Pythia’s, supporting the Universality Hypothesis that core semantic representations are similar across architectures. A tiny 0.02% of features diverge, with Mamba’s recurrent bottleneck causing it to compress syntactic anomalies into polysemantic neurons, whereas Pythia’s attention can isolate distinct formatting edge‑cases.
Sparse autoencoders (SAEs) are proposed to extract numerous features from large language model (LLM) representations, yet explaining these features still relies primarily on external observation. This reliance leads to superficial explanations inferred from observed model behavior and computational inefficiency from collecting such behavioral evidence at scale.
arXiv:2610.20139v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) have emerged as a promising approach to mechanistic interpretability by decomposing language-model activations into sparse,...
arXiv:2602. 10352v2 Announce Type: replace-cross Abstract: Self-interpretation methods prompt language models to describe their own internal states, but remain unreliable due to hyperparameter sensitivity.
The study investigates whether attention weights in Vision‑Language Models (VLMs) accurately reflect model reasoning for visual inputs. Using causal perturbation analysis, it identifies three distinct processing modes—Faithful‑Sufficient, Faithful‑Distributed, and Non‑Focal—indicating heterogeneous visual attention faithfulness. The research also shows that human‑annotated ground‑truth regions align with model attention in only about 60% of cases, highlighting a systematic divergence between model visual reliance and human intuition across VQA, document, and chart tasks.