Français(FR) Are Causal Neuron Circuits Unique and Transferable?

研究质疑LLM因果神经元回路的独特性和可迁移性

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 05:00

研究人员调查了大语言模型中因果神经元回路的必要性和可迁移性。他们的研究发现，虽然特定的神经元组会影响模型的输出，但通常只需要这些神经元的一个子集就能达到效果。研究还表明，这些因果效应通常可以迁移到改写后的输入，但在迁移到不同的推理链时表现出较弱的可迁移性。 AI

影响探讨了LLM中学习到的表征的基本性质，可能影响未来的可解释性研究和模型调试。

排序理由该集群包含一篇详细介绍LLM可解释性研究结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 33 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍LLM可解释性研究结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

arXiv cs.CL TIER_1 Français(FR) · \"Omer Tekin Yavuz (\.Ileti\c{s}im Yaz{\i}l{\i}m) · 2026-11-11 05:00

因果神经元回路是独特的且可迁移的吗？

arXiv:2611.13885v1 Announce Type: cross Abstract: Interpretability demonstrations often show that intervening on a group of neurons changes a model's answer. A recent example removes an 18-neuron group in Qwen3-32B and the model answers "Simba" instead of "Scar" to a two-hop Lion…