Perturbation Probing: A Two-Pass-per-Prompt Diagnostic for FFN Behavioral Circuits in Aligned LLMs
Perturbation probing generates task-specific causal hypotheses for FFN neurons in large language models using two forward passes per prompt and no backpropagation, followed by a one-time intervention sweep of about 150 passes amortized across all identified neurons. Across eight behavioral circuits, 13 models, and four architecture families, we identify two circuit structures that organize LLM behavior. Opposition circuits appear when RLHF suppresses a pre-training tendency. In safety refusal, about 50 neurons, or 0.014 percent of all neurons, control the refusal template; ablating them changes 80 percent of response formats on 520 AdvBench prompts while producing near-zero harmful compli...
| Authors | Hongliang Liu, Tung-Ling Li, Yuhao Wu |
|---|---|
| Primary category | cs.CL |
| Published | 30 April 2026 |
| Last updated at source | 30 April 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.