PRISM: A Dual View of LLM Reasoning through Semantic Flow and Latent Computation

• URL: http://arxiv.org/abs/2603.22754v1
• Date: Tue, 24 Mar 2026 03:31:53 GMT
• Title: PRISM: A Dual View of LLM Reasoning through Semantic Flow and Latent Computation
• Abstract summary: PRISM (Probabilistic Reasoning Inspection through Semantic and Implicit Modeling) is a framework and diagnostic tool for jointly analyzing both levels.<n>It uncovers systematic patterns in the reasoning process, showing that failed trajectories are more likely to become trapped in unproductive verification loops.<n> PRISM makes these behaviors observable and analyzable rather than relying solely on final-task accuracy.
• Score: 15.91920027845529
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) solve complex problems by generating multi-step reasoning traces. Yet these traces are typically analyzed from only one of two perspectives: the sequence of tokens across different reasoning steps in the generated text, or the hidden-state vectors across model layers within one step. We introduce PRISM (Probabilistic Reasoning Inspection through Semantic and Implicit Modeling), a framework and diagnostic tool for jointly analyzing both levels, providing a unified view of how reasoning evolves across steps and layers. Across multiple reasoning models and benchmarks, PRISM uncovers systematic patterns in the reasoning process, showing that failed trajectories are more likely to become trapped in unproductive verification loops and further diverge into distinct modes such as overthinking and premature commitment, which behave differently once a candidate answer is reached. It further reveals how prompting reshapes reasoning behavior beyond aggregate accuracy by altering both semantic transitions and internal computational patterns. By modeling reasoning trajectories as structured processes, PRISM makes these behaviors observable and analyzable rather than relying solely on final-task accuracy. Taken together, these insights position PRISM as a practical tool for analyzing and diagnosing reasoning processes in LLMs.

Related papers

• Revealing Algorithmic Deductive Circuits for Logical Reasoning [4.839520296557773]
  Large Language Models (LLMs) can achieve strong reasoning performance by incorporating functional functional symbolic representations.<n>This work aims to localize the attention heads responsible for individual reasoning steps and characterize the types of information transferred among them.
  arXiv  Detail & Related papers  (2026-05-27T01:30:09Z)
• Implicit Patterns in LLM-Based Binary Analysis [6.612857121548629]
  We present the first large-scale, trace-level study showing that multi-pass LLM reasoning gives rise to structured, token-level implicit patterns.<n>Our results provide the first systematic characterization of LLM-driven binary analysis and a foundation for more reliable analysis systems.
  arXiv  Detail & Related papers  (2026-03-19T16:56:56Z)
• Dynamics Within Latent Chain-of-Thought: An Empirical Study of Causal Structure [58.89643769707751]
  We study latent chain-of-thought as a manipulable causal process in representation space.<n>We find that latent-step budgets behave less like homogeneous extra depth and more like staged functionality with non-local routing.<n>These results motivate mode-conditional and stability-aware analyses as more reliable tools for interpreting and improving latent reasoning systems.
  arXiv  Detail & Related papers  (2026-02-09T15:25:12Z)
• Process In-Context Learning: Enhancing Mathematical Reasoning via Dynamic Demonstration Insertion [11.708864769915857]
  We propose Process In-Context Learning (PICL) to boost mathematical reasoning by responding to real-time inference needs.<n>PICL operates in two stages: 1)it identifies potential confusion points by analyzing semantics and entropy in the reasoning process and summarizes their core characteristics.<n>It retrieves relevant demonstrations from the demonstration pool that match the confusion context and inserts them directly into the ongoing reasoning process to guide subsequent steps.
  arXiv  Detail & Related papers  (2026-01-17T09:20:06Z)
• How and Why LLMs Generalize: A Fine-Grained Analysis of LLM Reasoning from Cognitive Behaviors to Low-Level Patterns [51.02752099869218]
  Large Language Models (LLMs) display strikingly different generalization behaviors.<n>We introduce a novel benchmark that decomposes reasoning into atomic core skills.<n>We show that RL-tuned models maintain more stable behavioral profiles and resist collapse in reasoning skills, whereas SFT models exhibit sharper drift and overfit to surface patterns.
  arXiv  Detail & Related papers  (2025-12-30T08:16:20Z)
• Schoenfeld's Anatomy of Mathematical Reasoning by Language Models [56.656180566692946]
  We adopt Schoenfeld's Episode Theory as an inductive, intermediate-scale lens and introduce ThinkARM (Anatomy of Reasoning in Models)<n>ThinkARM explicitly abstracts reasoning traces into functional reasoning steps such as Analysis, Explore, Implement, verify, etc.<n>We show that episode-level representations make reasoning steps explicit, enabling systematic analysis of how reasoning is structured, stabilized, and altered in modern language models.
  arXiv  Detail & Related papers  (2025-12-23T02:44:25Z)
• PRISM-Bench: A Benchmark of Puzzle-Based Visual Tasks with CoT Error Detection [33.98107848014898]
  We introduce PRISM-Bench, a benchmark of puzzle-based visual challenges designed to evaluate multimodal large language models (MLLMs)<n>Unlike prior evaluations that measure only final-answer accuracy, PRISM-Bench introduces a diagnostic task: given a visual puzzle and a step-by-step chain-of-thought containing exactly one error, models must identify the first incorrect step.<n>The puzzles in PRISM-Bench require multi-step symbolic, geometric, and analogical reasoning, resisting shortcuts based on superficial pattern matching.
  arXiv  Detail & Related papers  (2025-10-27T17:57:52Z)
• Implicit Reasoning in Large Language Models: A Comprehensive Survey [67.53966514728383]
  Large Language Models (LLMs) have demonstrated strong generalization across a wide range of tasks.<n>Recent studies have shifted attention from explicit chain-of-thought prompting toward implicit reasoning.<n>This survey introduces a taxonomy centered on execution paradigms, shifting the focus from representational forms to computational strategies.
  arXiv  Detail & Related papers  (2025-09-02T14:16:02Z)
• CLATTER: Comprehensive Entailment Reasoning for Hallucination Detection [60.98964268961243]
  We propose that guiding models to perform a systematic and comprehensive reasoning process allows models to execute much finer-grained and accurate entailment decisions.<n>We define a 3-step reasoning process, consisting of (i) claim decomposition, (ii) sub-claim attribution and entailment classification, and (iii) aggregated classification, showing that such guided reasoning indeed yields improved hallucination detection.
  arXiv  Detail & Related papers  (2025-06-05T17:02:52Z)
• How do Transformers Learn Implicit Reasoning? [67.02072851088637]
  We study how implicit multi-hop reasoning emerges by training transformers from scratch in a controlled symbolic environment.<n>We find that training with atomic triples is not necessary but accelerates learning, and that second-hop generalization relies on query-level exposure to specific compositional structures.
  arXiv  Detail & Related papers  (2025-05-29T17:02:49Z)
• Distributional reasoning in LLMs: Parallel reasoning processes in multi-hop reasoning [8.609587510471943]
  We introduce a novel and interpretable analysis of internal multi-hop reasoning processes in large language models. We show that during inference, the middle layers of the network generate highly interpretable embeddings. Our findings can help uncover the strategies that LLMs use to solve reasoning tasks, offering insights into the types of thought processes that can emerge from artificial intelligence.
  arXiv  Detail & Related papers  (2024-06-19T21:36:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.