finding
active
finding:finetuning-primarily-makes-shallow-edits-to-existing-representations-and-circuits-so-harmlessness-training-is-cheaply-undone

Finetuning primarily makes shallow edits to existing representations and circuits, so harmlessness training is cheaply undone

Supported by further finetuning and causal interventions.

Source paper

extracted_from
Open Problems in Mechanistic Interpretability
(2025) · Lee Sharkey · Bilal I. Chughtai · Joshua Batson · Jack W. Lindsey +25

Neighborhood — ranked by edge-count

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.