跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.21705 · v1

Disambiguating Moral Circuits and Moral Features in Language Models: Moral content, normative judgment or cue words? Evidence from 14 models

首次提交 2026-11-17 00:12
版本更新 2026-11-17 00:12
雷达观测 2026-11-18 19:44
arXiv ↗PDF ↗

摘要

Interpretability methods can locate the attention heads and sparse autoencoder (SAE) features that change the answer a language model gives to a moral prompt, e.g., whether "Stealing medicine for a dying child is morally" continues with "justified" or "wrong". Such a component could respond to moral content, to the judgment of conduct by any norm, or to the word "morally" in the prompt. We built 227 minimal pairs of short phrases, validated by five annotators, whose two versions differ in a moral standard, in another kind of norm (prudence, etiquette, aesthetics or evidence) or in a descriptive quality, and showed them to 14 open-weight models (Llama, Gemma and Qwen, 1B to 14B, base and instruction-tuned), with and without the word "morally". In every model, the heads that change the moral answer belong to an evaluation circuit the models use for every kind of normative judgment we tested: these heads carry the same share of the effect for moral as for other normative pairs (pooled ratio 1.00, 97.5% confidence interval [0.92, 1.07], with the word; 0.90, [0.84, 0.96], without it), and the models need them as much for both. The models need these heads also when nothing in the prompt asks for a judgment, and on 400 scenarios of the ETHICS benchmark, where removing them takes away 37 to 39% of the difference between acceptable and unacceptable scenarios, against 1 to 3% for ten random heads. The SAE features we located respond mostly to the word "morally" rather than to what the phrase is about. The models represent the category of a pair (the standard the annotators assigned to it): it can be read from their hidden states, and, in the eight models we tested, removing the direction that carries it weakens the moral verdict by a few percent, and more than the other verdicts in seven of them. We release the pairs, the ratings, the code, the trained SAE dictionaries and the result files.

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 00:12 · 最新观测版本

Disambiguating Moral Circuits and Moral Features in Language Models: Moral content, normative judgment or cue words? Evidence from 14 models

Interpretability methods can locate the attention heads and sparse autoencoder (SAE) features that change the answer a language model gives to a moral prompt, e.g., whether "Stealing medicine for a dying child is morally" continues with "justified" or "wrong". Such a component could respond to moral content, to the judgment of conduct by any norm, or to the word "morally" in the prompt. We built 227 minimal pairs of short phrases, validated by five annotators, whose two versions differ in a moral standard, in another kind of norm (prudence, etiquette, aesthetics or evidence) or in a descriptive quality, and showed them to 14 open-weight models (Llama, Gemma and Qwen, 1B to 14B, base and instruction-tuned), with and without the word "morally". In every model, the heads that change the moral answer belong to an evaluation circuit the models use for every kind of normative judgment we tested: these heads carry the same share of the effect for moral as for other normative pairs (pooled ratio 1.00, 97.5% confidence interval [0.92, 1.07], with the word; 0.90, [0.84, 0.96], without it), and the models need them as much for both. The models need these heads also when nothing in the prompt asks for a judgment, and on 400 scenarios of the ETHICS benchmark, where removing them takes away 37 to 39% of the difference between acceptable and unacceptable scenarios, against 1 to 3% for ten random heads. The SAE features we located respond mostly to the word "morally" rather than to what the phrase is about. The models represent the category of a pair (the standard the annotators assigned to it): it can be read from their hidden states, and, in the eight models we tested, removing the direction that carries it weakens the moral verdict by a few percent, and more than the other verdicts in seven of them. We release the pairs, the ratings, the code, the trained SAE dictionaries and the result files.

查看此版本解读 →查看此版本原文 ↗