#mechanistic interpretability

14 stories · 2 in 7 days · 8 in 30 days
Stories, products and papers tagged "mechanistic interpretability", newest first.
Nov 18
IndustryUS vs ChinaMulti-source71°00:36
Chinese contentOpenAI 收购 Goodfire 谈判期间,Sam Altman 特别关注研究员 Nick Cammarata

The Information 的独家爆料,细节很鲜:OpenAI 想买可解释性公司 Goodfire,Altman 点名欣赏前员工 Cammarata 的文章。想了解大厂怎么抢可解释性人才、这领域为何吃香的话,这篇值得看看。

Event
The Information@theinformationSource
Nov 16
PapersBuildersMulti-source13:28
Chinese content机制可解释性研究解释 Web 智能体的幻觉失败:线性探针达 0.80 AUC

如果你做过浏览器操作智能体就会懂:它老点不存在的按钮。这篇用 Qwen2.5-7B-Instruct 做了很扎实的机制分析,探针预测 0.80 AUC、steering 纠 85% 幻觉,但格式一变就失效,把可解释性方法的局限讲得很诚实。做智能体评测和可解释性研究的人都该看看。

Event
arXiv cs.AI@Mohamed Aghzal et al. (4)Source
Nov 10
Nov 7
Nov 6
PapersBuildersSelected13:40
Survey Maps Mechanistic Interpretability Methods for AI Safety

The survey's value lies in its proposed evaluation framework and its six-way taxonomy of research orientations, which gives structure to a field that has grown mostly through scattered techniques. Researchers and safety teams can use the released notebooks to reproduce methods hands-on rather than reading only about them. The comparison between mechanistic evidence and behavioural evaluation is a useful lens for anyone deciding whether to invest in interpretability tools over standard safety testing.

arXiv cs.LG@Fazl Barez et al. (27)Source
Nov 2
Oct 26
Oct 23
Oct 16
IndustrySelected15:32
Polysemanticity Death Claim Questioned in Interpretability Debate

The argument pushes back on a popular narrative in mechanistic interpretability rather than adding new data, reminding readers that monosemanticity assumptions may not survive contact with real networks. Researchers and safety practitioners evaluating SAE-style interpretability pipelines should weigh this critique when deciding how much to trust neuron-level explanations. No regional angle applies; the debate is methodological.

Teortaxes@teortaxesTexSource
PapersBuildersMulti-sourceSelected02:36
Nick Cammarata says most neurons in trillion-parameter models are legible

The notable shift is methodological: scale plus automated agents, not new theory, appears to have cracked neuron legibility, echoing the bitter-lesson pattern from other AI fields. Interpretability researchers and alignment teams should test these automated pipelines on their own models rather than assuming internals remain opaque. The claims come from a practitioner essay rather than a peer-reviewed paper, so treat the quantitative picture as one lab's report.

Event
elvis@omarsar0Source
Aug 28
Jul 14
Jun 16