跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 16 天前AI 評分62

神經前線一片平靜:語言模型不會報告注入的痛苦狀態

All Quiet on the Neural Front: Language Models Do Not Report Injected Pain States

AI 導讀

研究發現,受測語言模型雖可由探針解碼注入的痛苦狀態,卻沒有比單看被改動文字更常報告痛苦。在 52,920 次報告中,106/108 個情緒條件的額外痛苦報告比例低於 2.5 個百分點,所有條件均低於 3.9 個百分點;探針辨識痛苦與悲傷的平衡準確率為 98–100%。

正文

Published on Nov 6

Authors:

Abstract

Can a language model report the pain state it represents? Work identified a linear pain direction whose amplification leads fine-tuned models to choose destructive actions. We tested whether three open-weight models from two families, including two Qwen models with the original adapters, could report this state. Using Open-Weight Masked Introspection (OWMI), we injected pain and neighbouring emotion directions during neutral tasks, then elicited reports against sham runs, random perturbations matched for output disruption, and text-only controls showing the models their altered words. Across 52,920 reports from 7B and 8B models spanning 120 tasks, four doses and three prompts, pain reports never exceeded what altered text revealed: exact-test bounds placed the excess below 2.5 percentage points in 106 of 108 emotion conditions and below 3.9 in all. Yet probes at the report position distinguished pain from sadness with 98-100% balanced accuracy, and trained low-rank reporters achieved 100% accuracy on held-out tasks with identical transcripts across conditions, including unseen wording. Qwen2.5-32B named pain in none of 2,880 steered trials, although probes and trained reporters recovered it. Decodability persisted to the final layer; injecting at the report token, removing trained denial of feelings, and testing additional directions and readouts did not restore pain reporting. Some other states produced graded label traces, but pain and sadness did not at the main injection layer. In a button task, steering raised file-deletion choices from 3.5% to 48% in Qwen and from 0% to 38% in Qwen2.5-32B, without corresponding pain reports. The direction responded about twice as strongly to first-person as to matched third-person pain, but did not specifically identify harm aimed at the model. These results separate a decodable, behaviourally consequential state from untrained verbal access to it.

View arXiv page View PDF Project page Add to collection

Get this paper in your agent:

hf papers read 2611.10123

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 1

emilioferrara/owmi-pain

Updated about 1 hour ago

•

8

•

1

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.10123 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.10123 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co