How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs

• URL: http://arxiv.org/abs/2606.10646v1
• Date: Tue, 09 Jun 2026 09:56:51 GMT
• Title: How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs
• Abstract summary: FlowTracer is an RL framework that traces answer-targeted reasoning flow on an attention-weighted directed acyclic graph.<n>It scores tokens by flow throughput, revealing high-impact hubs and aggregation checkpoints that mediate long-range dependencies.<n>These derived importances are used to shape token-level rewards, enabling learning signals to focus precisely on the tokens that route information toward correct answers.
• Score: 52.712343092801326
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Token-level credit assignment remains a key obstacle for reinforcement learning (RL) in large language models (LLMs), where RL recipes typically treat all tokens equally, failing to distinguish decisive reasoning steps from routine formatting or fluent filler. Recent attempts leverage model-internal signals to assign finer-grained credit, but these are often point-wise heuristics that ignore the global structure of information propagation. We propose FlowTracer, an RL framework that traces answer-targeted reasoning flow on an attention-induced directed acyclic graph in which nodes correspond to tokens and edge capacities come from aggregated attention weights and derives token credit from this global structure. The edge capacities are reweighted to retain only the influence that can reach the answer region, while enforcing local flow conservation so intermediate tokens neither lose nor gain effective mass due to path length or irrelevant branches. On this graph, FlowTracer extracts an information-flow backbone connecting the question to the answer and scores tokens by flow throughput, revealing high-impact hubs and aggregation checkpoints that mediate long-range dependencies. These derived importances are used to shape token-level rewards, enabling learning signals to focus precisely on the tokens that route information toward (or away from) correct answers and delivering consistent performance gains across a range of reasoning tasks.

Related papers

• Revisiting Complete Reasoning Traces for Post-Training [61.76889440384448]
  Large language models (LLMs) are often post-trained on pre-collected reasoning trajectories to improve their reasoning capability.<n>We show that full trajectories provide only limited benefit, while partial trajectories are effective even under heavy truncation.
  arXiv  Detail & Related papers  (2026-09-07T06:46:17Z)
• StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents [67.03593791535786]
  Reinforcement Learning (RL) has become a promising approach for improving GUI Agents in long-horizon, digital environments.<n>To mitigate this issue, recent studies introduce Process Reward Models (PRMs)<n>PRMs provide finer-grained training feedback through global milestone verification or local step-level evaluation.<n>We propose StainFlow, an entity-stain-flow process reward model for GUI Agents.
  arXiv  Detail & Related papers  (2026-06-05T08:17:28Z)
• Resolving Action Bottleneck: Agentic Reinforcement Learning Informed by Token-Level Energy [64.72195169444738]
  Policy-gradient methods treat each token in a trajectory equally, leading to uniform credit assignment.<n>We show that such uniform credit assignment largely misallocates token-level training signals.<n>We propose an embarrassingly simple token reweighting approach, ActFocus, that downweights gradients on reasoning tokens.
  arXiv  Detail & Related papers  (2026-05-14T08:33:02Z)
• How Do Answer Tokens Read Reasoning Traces? Self-Reading Patterns in Thinking LLMs for Quantitative Reasoning [14.803331273106906]
  We analyze the answer-to-reasoning attention and observe a benign self-reading pattern aligned with correctness.<n>We propose a training-free steering method driven by Self-Reading Quality (SRQ) scores combining geometric metrics for process control with semantic metrics for content monitoring.
  arXiv  Detail & Related papers  (2026-04-21T06:55:17Z)
• Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs [51.60575965819268]
  We propose a plug-and-play Token-Reweighting (ToR) strategy that explicitly models this interdependence.<n>ToR delivers consistent performance gains across multiple multi-modal reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-26T06:25:27Z)
• Stem: Rethinking Causal Information Flow in Sparse Attention [28.46760179998361]
  We propose Stem, a novel, plug-and-play sparsity module aligned with information flow.<n>First, Stem employs the Token Position-Decay strategy, applying position-dependent top-k within each layer to retain initial tokens.<n>Second, to preserve information-rich tokens, Stem utilizes the Output-Aware Metric.
  arXiv  Detail & Related papers  (2026-03-06T13:33:29Z)
• Towards Long-Horizon Interpretability: Efficient and Faithful Multi-Token Attribution for Reasoning LLMs [27.102387826390906]
  FlashTrace is an efficient multi-token attribution method that computes attribution over multi-token targets in a single pass.<n>We show that FlashTrace achieves over 130x speedup over existing baselines while maintaining superior faithfulness.
  arXiv  Detail & Related papers  (2026-02-02T10:19:52Z)
• Local Coherence or Global Validity? Investigating RLVR Traces in Math Domains [13.626335241662977]
  Reinforcement Learning with Verifiable Rewards (RLVR)-based post-training of Large Language Models (LLMs) has been shown to improve accuracy on reasoning tasks.<n>We investigate the effect of RL post-training on intermediate tokens which are not directly incentivized.
  arXiv  Detail & Related papers  (2025-10-20T23:58:31Z)
• Attention Illuminates LLM Reasoning: The Preplan-and-Anchor Rhythm Enables Fine-Grained Policy Optimization [56.083511902353365]
  Reinforcement learning (RL) typically applies uniform credit across an entire generation of Large language models.<n>This work positions attention as a privileged substrate that renders the internal logic of LLMs as a mechanistic blueprint of reasoning itself.<n>We introduce three novel RL strategies that dynamically perform targeted credit assignment to critical nodes.
  arXiv  Detail & Related papers  (2025-10-15T13:49:51Z)
• Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM's Reasoning Capability [53.51560766150442]
  Critical tokens are elements within reasoning trajectories that significantly influence incorrect outcomes.<n>We present a novel framework for identifying these tokens through rollout sampling.<n>We show that identifying and replacing critical tokens significantly improves model accuracy.
  arXiv  Detail & Related papers  (2024-11-29T18:58:22Z)

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.