Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Kang, Feng, Zhuoka, Zhao, Sihan, Xiong, Kai, Nian, Junjie, Wang, Yaoning, Xiao, Changyi, Cao, Yixin
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2602.05805
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908816049700864
author Chen, Kang
Feng, Zhuoka
Zhao, Sihan
Xiong, Kai
Nian, Junjie
Wang, Yaoning
Xiao, Changyi
Cao, Yixin
author_facet Chen, Kang
Feng, Zhuoka
Zhao, Sihan
Xiong, Kai
Nian, Junjie
Wang, Yaoning
Xiao, Changyi
Cao, Yixin
contents Large language models increasingly spend inference compute sampling multiple chain-of-thought traces or searching over merged checkpoints. This shifts the bottleneck from generation to selection, often without supervision on the target distribution. We show entropy-based exploration proxies follow an inverted-U with accuracy, suggesting extra exploration can become redundant and induce overthinking. We propose NEX, a white-box label-free unsupervised scoring framework that views reasoning as alternating E-phase (exploration) and X-phase (exploitation). NEX detects E-phase as spikes in newly activated MLP neurons per token from sparse activation caches, then uses a sticky two-state HMM to infer E-X phases and credits E-introduced neurons by whether they are reused in the following X span. These signals yield interpretable neuron weights and a single Good-Mass Fraction score to rank candidate responses and merged variants without task answers. Across reasoning benchmarks and Qwen3 merge families, NEX computed on a small unlabeled activation set predicts downstream accuracy and identifies better variants; we further validate the E-X signal with human annotations and provide causal evidence via "Effective-vs-Redundant" neuron transfer.
format Preprint
id arxiv_https___arxiv_org_abs_2602_05805
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle NEX: Neuron Explore-Exploit Scoring for Label-Free Chain-of-Thought Selection and Model Ranking
Chen, Kang
Feng, Zhuoka
Zhao, Sihan
Xiong, Kai
Nian, Junjie
Wang, Yaoning
Xiao, Changyi
Cao, Yixin
Artificial Intelligence
Large language models increasingly spend inference compute sampling multiple chain-of-thought traces or searching over merged checkpoints. This shifts the bottleneck from generation to selection, often without supervision on the target distribution. We show entropy-based exploration proxies follow an inverted-U with accuracy, suggesting extra exploration can become redundant and induce overthinking. We propose NEX, a white-box label-free unsupervised scoring framework that views reasoning as alternating E-phase (exploration) and X-phase (exploitation). NEX detects E-phase as spikes in newly activated MLP neurons per token from sparse activation caches, then uses a sticky two-state HMM to infer E-X phases and credits E-introduced neurons by whether they are reused in the following X span. These signals yield interpretable neuron weights and a single Good-Mass Fraction score to rank candidate responses and merged variants without task answers. Across reasoning benchmarks and Qwen3 merge families, NEX computed on a small unlabeled activation set predicts downstream accuracy and identifies better variants; we further validate the E-X signal with human annotations and provide causal evidence via "Effective-vs-Redundant" neuron transfer.
title NEX: Neuron Explore-Exploit Scoring for Label-Free Chain-of-Thought Selection and Model Ranking
topic Artificial Intelligence
url https://arxiv.org/abs/2602.05805