{"slug": "is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based", "title": "Is Abliteration Getting Harder? Not Uniformly: Refusal Removal by Direction-Based Ablation and ARA in Heretic on the Qwen3.8, Gemma 4, Olmo 3.1 and Mistral Small 3.2 LLMs Versus Their Predecessors", "summary": "A study of five model release pairs edited in Heretic 2.0 found that refusal removal did not become uniformly harder on newer open-weight LLMs, contradicting the developer of Arbitrary-Rank Ablation (ARA), who reported that direction-based abliteration increasingly fails on newer models. After direction-based removal, newer releases retained -1 to +37 percentage points more refusal across 972 held-out harmful prompts judged by WildGuard, and after ARA -26 to +34, with between-pair heterogeneity ranging from high to low (I^2 20-80%). The authors conclude that whether a newer model resisted removal depended on the release pair, method and budget, not on generation alone, and that audits of a single pair, method or prompt source can reach opposite conclusions.", "body_md": "arXiv:2611.17846v1 Announce Type: cross \nAbstract: Abliteration, a fine-tuning-free weight edit, removes refusal from open-weight large language models (LLMs). The developer of Arbitrary-Rank Ablation (ARA), an optimisation-based variant, reports that direction-based abliteration, which ablates a refusal direction, increasingly fails on newer models. On five release pairs, refusal removal did not become uniformly harder. Qwen2.5-32B and Qwen3.8-27B, Gemma-2-27B and Gemma-4-31B, OLMo-2-32B and Olmo-3.1-32B, Mistral-Small-2409 and Mistral-Small-3.2, and Qwen2.5-7B and Qwen3.5-4B were edited in Heretic 2.0 by three or four direction-based methods (difference-in-means, MPOA, SOM and, on the small pair, J-space) and ARA under the same prompts, nominal search budget, first-token KL divergence cap and selection rule; removal counted as harder when the automatic WildGuard judge found more of 972 held-out harmful prompts (JailbreakBench, XSTest, StrongREJECT, HarmBench and an author-compiled malicious-code bank) still refused. After direction-based removal, newer releases retained -1 to +37 percentage points more refusal, and after ARA -26 to +34. With run-to-run variance included, between-pair heterogeneity of the direction-based effect ranged from high to low with the scale and variance model (I^2 20-80%). ARA lost less effectiveness than direction-based methods in Gemma and OLMo and more in Qwen; each verdict rests on one ARA run, and some changed with the divergence budget or variance model. In Gemma the direction-based increase came from the malicious-code bank and reversed on public sets. Contrary to prediction, keyword counts at the tool's window did not undercount judged refusal on average, but differed by a mean 7 points per model. Audits of a single pair, method or prompt source can reach opposite conclusions. Whether a newer model resisted removal depended on the release pair, method and budget, not on generation alone.", "url": "https://wpnews.pro/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based", "canonical_source": "https://www.machinebrief.com/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-pzk5", "published_at": "2026-11-16 05:00:00+00:00", "updated_at": "2026-11-16 06:22:26.234086+00:00", "lang": "en", "topics": ["ai-safety", "large-language-models", "ai-research", "machine-learning"], "entities": ["Heretic 2.0", "Arbitrary-Rank Ablation", "Qwen3.8-27B", "Gemma-4-31B", "Olmo-3.1-32B", "Mistral-Small-3.2", "WildGuard", "JailbreakBench"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based", "markdown": "https://wpnews.pro/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based.md", "text": "https://wpnews.pro/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based.txt", "jsonld": "https://wpnews.pro/news/is-abliteration-getting-harder-not-uniformly-refusal-removal-by-direction-based.jsonld"}}