REVIEW 7 cited by
Refusal in LLMs is an Affine Function
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We propose affine concept editing (ACE) as an approach for steering language models' behavior by intervening directly in activations. We begin with an affine decomposition of model activation vectors and show that prior methods for steering model behavior correspond to subsets of terms of this decomposition. We then provide a derivation of ACE and use it to control refusal behavior on ten different models, including Llama 3 70B. ACE combines affine subspace projection and activation addition to reliably control the model's refusal responses across prompt types. We evaluate the results using LLM-based scoring on a collection of harmful and harmless prompts. Our experiments demonstrate that ACE consistently achieves more precise control over model behavior than existing methods and generalizes to models where directional ablation via affine subspace projection alone produces incoherent outputs. Code for reproducing our results is available at https://github.com/EleutherAI/steering-llama3 .
Forward citations
Cited by 7 Pith papers
-
Low-Resource Safety Failures Are Action Failures, Not Representation Failures
Low-resource safety failures are action failures because the harmfulness representation transfers but the decision calibration does not; this is fixed by recalibrating a high-resource gate with 1-4 target-language examples.
-
Less is More: Geometric Unlearning for LLMs with Minimal Data Disclosure
Geometric Unlearning suppresses specific knowledge in LLMs by projecting hidden planning states onto a low-rank safe geometry derived from minimal reference prompts.
-
Expert-Aware Refusal Steering
Refusal steering works on MoE LLMs; expert-aware variants succeed with single-expert outputs and refusal signals differ from routing patterns.
-
RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMs
RefusalGuard constrains updates in hidden representation space to preserve safety-relevant geometric structure during fine-tuning, maintaining low attack success rates on safety benchmarks while preserving task performance.
-
Less is More: Geometric Unlearning for LLMs with Minimal Data Disclosure
Geometric Unlearning distills a low-rank safe subspace from reference prompts and applies projection-based alignment on synthetic anchors to suppress target content while preserving non-target utility.
-
Activation Steering for Accent Adaptation in Large Audio Language Models
Mean-shift activation steering in middle encoder layers of Qwen2-Audio reduces ASR word error rates across eight accents without updating model parameters.
-
Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications
Empirical comparison of alignment ablation methods on a 60-prompt security evaluation suite shows task-only LoRA achieves 0.87 mean security score with 0.13 unsafe compliance.
Discussion (0). Sign in to comment.