SPREG: Structured Plan Repair with Entropy-Guided Test-Time Intervention for Large Language Model Reasoning

2026-04-20 · Xuan Wang, Yu Ming, Xinhao Zhong, Xinyu Yu, Wenjie Wang, Shuai Chen, Wei Lin arxiv

Large Language Models (LLMs) are prone to logical hallucinations and stochastic drifts during long-chain reasoning. While Classifier-Free Guidance (CFG) can improve instruction adherence, standard static implementations often cause semantic dilution and linguistic degradation. We propose SPREG (Structured Plan-guided Real-time Entropy Gating), a lightweight inference-time framework for surgical error rectification. SPREG employs an adaptive dual-threshold mechanism to monitor real-time entropy, identifying sudden ``entropy spikes'' as reliable indicators of logical failure. Upon detection, it triggers a dynamic repair by replacing uninformative null-priors with reference distributions synthesized from historical high-confidence states. By modulating guidance intensity according to structured reasoning stages (e.g., Action, Observation), SPREG steers the model back to a stable manifold without compromising fluency. Our experiments demonstrate significant gains, notably a 20.0% absolute accuracy improvement on AIME25, while effectively suppressing uncontrolled entropy drift in complex tasks.

📄 PDF Abstract BibTeX arXiv:2604.17884

Code (0)

등록된 구현이 없습니다.

ScenePilot: Grow-and-Repair Policy for Text-Driven 3D Indoor Scene Generation

2026-08-31 · Jiawei Zhang, Hongsong Wang, Pan Zhou arxiv

Text-driven 3D indoor scene generation has advanced from dataset-bound layout modeling to open-vocabulary synthesis with large language and vision-language models. Yet existing methods remain limited: one-pass generators…

Scene Generation

Structural Entropy Guided Agent for Detecting and Repairing Knowledge Deficiencies in LLMs

2025-05-12 · Yifan Wei, Xiaoyan Yu, Tengfei Pan, Angsheng Li 외

Large language models (LLMs) have achieved unprecedented performance by leveraging vast pretraining corpora, yet their performance remains suboptimal in knowledge-intensive domains such as medicine and scientific researc…

AI Agent Knowledge Distillation Knowledge Graphs Reinforcement Learning+1

HEAL: Hindsight Entropy-Assisted Learning for Reasoning Distillation

2026-03-11 · Wenjing Zhang, Jiangze Yan, Jieyun Huang, Yi Shen 외 arxiv

Distilling reasoning capabilities from Large Reasoning Models (LRMs) into smaller models is typically constrained by the limitation of rejection sampling. Standard methods treat the teacher as a static filter, discarding…

Vibe Coding an LLM-powered Theorem Prover

2026-01-08 · Zhe Hou arxiv

We present Isabellm, an LLM-powered theorem prover for Isabelle/HOL that performs fully automatic proof synthesis. Isabellm works with any local LLM on Ollama and APIs such as Gemini CLI, and it is designed to run on con…

Code Generation

VISTA: Knowledge-Driven Vessel Trajectory Imputation with Repair Provenance

2026-01-11 · Hengyu Liu, Tianyi Li, Haoyu Wang, Kristian Torp 외 arxiv

Repairing incomplete trajectory data is essential for downstream spatio-temporal applications. Yet, existing repair methods focus solely on reconstruction without documenting the reasoning behind repair decisions, underm…

Anomaly Detection