跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.22234 · v1

The Fragility of Reasoning-Importance Attribution for Pruning Language Models

首次提交 2026-11-17 13:36
版本更新 2026-11-17 13:36
雷达观测 2026-11-18 19:43
arXiv ↗PDF ↗

摘要

Attribution-guided pruning scores components by their causal contribution and protects the important ones, and it is increasingly applied to compress reasoning language models without destroying their chain of thought. Its reliability rests on an unexamined assumption: that reasoning importance can be faithfully measured. Across four small reasoning models, we show this measurement is fragile in three ways, each of which changes pruning decisions. First, scoring final-answer tokens, the natural choice, saturates in every chain-of-thought model we test (answer-token negative log-likelihood, NLL, of 10−410^{-4}--10−310^{-3}) because the model states its answer inside its own reasoning; guiding head pruning with this target costs 27.6 and 32.2 accuracy points relative to scoring reasoning tokens on the two models we power with five seeds, while perplexity barely moves and even reverses sign at 7B. Second, causal importance is not localizable to single units: single-neuron effects sit at the measurement floor, their cross-split ranking is noise, and they cancel: a group's ablation effect disagrees in sign with the summed effects of its members in 90\% of groups. Reproducible importance emerges only at groups of ∼100\sim 100 neurons. Third, contrastively isolated ``reasoning-specific'' importance is not separable from generic importance, and protecting it is worse than protecting raw importance. We additionally pre-registered the hypothesis that distillation-induced concentration of importance predicts the pruning cost; it is not supported, and since our grid confounds concentration with scale, we report it as unresolved rather than reversed. The result is a short set of measurement rules: score reasoning tokens, attribute at coarse granularity, and evaluate with accuracy rather than perplexity.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-17 13:36 · 最新观测版本

The Fragility of Reasoning-Importance Attribution for Pruning Language Models

Attribution-guided pruning scores components by their causal contribution and protects the important ones, and it is increasingly applied to compress reasoning language models without destroying their chain of thought. Its reliability rests on an unexamined assumption: that reasoning importance can be faithfully measured. Across four small reasoning models, we show this measurement is fragile in three ways, each of which changes pruning decisions. First, scoring final-answer tokens, the natural choice, saturates in every chain-of-thought model we test (answer-token negative log-likelihood, NLL, of 10−410^{-4}--10−310^{-3}) because the model states its answer inside its own reasoning; guiding head pruning with this target costs 27.6 and 32.2 accuracy points relative to scoring reasoning tokens on the two models we power with five seeds, while perplexity barely moves and even reverses sign at 7B. Second, causal importance is not localizable to single units: single-neuron effects sit at the measurement floor, their cross-split ranking is noise, and they cancel: a group's ablation effect disagrees in sign with the summed effects of its members in 90\% of groups. Reproducible importance emerges only at groups of ∼100\sim 100 neurons. Third, contrastively isolated ``reasoning-specific'' importance is not separable from generic importance, and protecting it is worse than protecting raw importance. We additionally pre-registered the hypothesis that distillation-induced concentration of importance predicts the pruning cost; it is not supported, and since our grid confounds concentration with scale, we report it as unresolved rather than reversed. The result is a short set of measurement rules: score reasoning tokens, attribute at coarse granularity, and evaluate with accuracy rather than perplexity.

查看此版本解读 →查看此版本原文 ↗