The story’s score and claims

Comments on No. 80 of 143 · Goodfire · edition of Wednesday 11 November 2026

Neurons as an Interpretable Basis

Summary

This study argues that individual MLP neurons in Qwen3.6-35B-A3B are interpretable units of computation, not the intractable, polysemantic units often assumed. Using activation sparsification, circuit attribution, and agent-driven experiments, the authors report that ablating 92.2% of weak activations raises language-model loss by only 2.9%.

Comments

Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules