跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2609.23125 · v2

Perplexity Ranks Activation Quantisers but Understates What They Remove

首次提交 2026-09-20 00:51
版本更新 2026-11-18 07:52
雷达观测 2026-11-19 19:08
arXiv ↗PDF ↗

摘要

Four-bit activation quantisation of language models is usually tuned and compared by perplexity. We ask what that misses, and find three things. First, perplexity gets the order of quantisers right but the size of their damage wrong. Across four models it ranks 33 of 34 pairs of settings correctly by closed-book QA loss, yet a 6-16% rise in perplexity comes with a 10-23% drop in QA accuracy, and the lost facts are mostly rare ones. Second, the ranking only holds while the quantisation error acts like random noise. Without rotation, rounding small values to zero deletes part of the signal, and this hurts in-context retrieval far more than noise of the same size would. The crest factor of the quantiser's input (its peak-to-average ratio) helps tell the two cases apart. Third, rotating in blocks, the usual workaround when a layer's width has no ready-made Hadamard matrix, can land a model in between and leave outliers behind; 10 of 42 popular open models have such a width. A full-width Hadamard exists for every such width we found, costs about the same, and on Qwen3-4B halves the closed-book TriviaQA loss (0.112 to 0.060, over three rotation seeds). We recommend reporting a knowledge-heavy task alongside perplexity, and rotating at full width wherever blocks leave an input outlier-heavy.

作者备注:55 pages, 30 figures, 40 tables

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

本站首次收录为 v2,暂无更早版本可对照。

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。

v2 · 2026-11-18 07:52 · 最新观测版本

Perplexity Ranks Activation Quantisers but Understates What They Remove

Four-bit activation quantisation of language models is usually tuned and compared by perplexity. We ask what that misses, and find three things. First, perplexity gets the order of quantisers right but the size of their damage wrong. Across four models it ranks 33 of 34 pairs of settings correctly by closed-book QA loss, yet a 6-16% rise in perplexity comes with a 10-23% drop in QA accuracy, and the lost facts are mostly rare ones. Second, the ranking only holds while the quantisation error acts like random noise. Without rotation, rounding small values to zero deletes part of the signal, and this hurts in-context retrieval far more than noise of the same size would. The crest factor of the quantiser's input (its peak-to-average ratio) helps tell the two cases apart. Third, rotating in blocks, the usual workaround when a layer's width has no ready-made Hadamard matrix, can land a model in between and leave outliers behind; 10 of 42 popular open models have such a width. A full-width Hadamard exists for every such width we found, costs about the same, and on Qwen3-4B halves the closed-book TriviaQA loss (0.112 to 0.060, over three rotation seeds). We recommend reporting a knowledge-heavy task alongside perplexity, and rotating at full width wherever blocks leave an input outlier-heavy.

查看此版本解读 →查看此版本原文 ↗