Transformer circuits show free evolution during SFT, rendering static mechanistic localization inadequate for future parameter updates due to inherent temporal latency.
How does gpt-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model.Advances in Neural Information Processing Systems, 36:76033–76060, 2023
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Navigating by Old Maps: The Pitfalls of Static Mechanistic Localization in LLM Post-Training
Transformer circuits show free evolution during SFT, rendering static mechanistic localization inadequate for future parameter updates due to inherent temporal latency.