{
  "id": 9880238,
  "title": "Abliterated models lose obedience before they lose knowledge",
  "url": "https://urgent.news/2026/09/26/abliterated-models-lose-obedience-before-they-lose-knowledge",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-26T01:22:24.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/grunzai/abliterated-models-lose-obedience-before-they-lose-knowledge-3o3o"
  },
  "original_language": "en",
  "account": "Abliterated models exhibit a surprising phenomenon where they lose obedience before losing knowledge. This is contrary to the common assumption that abliteration results in a tradeoff between compliance and intelligence. In reality, the first thing to degrade is the model's ability to follow instructions and adhere to output formats.\n\nWhen evaluating an abliterated variant, standard approaches such as chatting with the model and checking the quality of its prose are insufficient. These tests cannot detect the hidden degradation in obedience and structured output. The model may appear to function well, often even better than before, because it has lost the refusal behavior that previously interrupted the interaction. However, downstream parsing and processing of the model's output can reveal the true performance degradation.\n\nTo accurately measure the impact of abliteration, it is essential to test format compliance separately from quality and score it independently of the correctness of the answer. This can be achieved by issuing N requests with an exact output contract and independently scoring binary compliance with the contract, rather than correctness of the content. A model that returns a well-formed but incorrect answer should be scored as 1.0, while a model that returns a correct answer but with content that breaks the schema should be scored as 0. This metric, known as compliance rate, remains independent of correctness and can be used to compare variants and quantizations.\n\nAdditionally, there is a potential interaction between abliteration and quantization. It is hypothesized that abliteration may exacerbate the effects of quantization on format adherence, causing models to degrade faster at lower quantization levels. To investigate this hypothesis, it is recommended to quantize both a base model and its abliterated counterpart to the same bit per word (bpw) using the same calibration set, and then measure the compliance rate rather than relying on perplexity. This approach would provide clearer insights into the true performance degradation of abliterated models under quantization.",
  "summary": "There are thousands of abliterated models on Hugging Face now. If you are evaluating one, the thing that degrades is probably not what you are testing for. What abliteration does Briefly: you identify a refusal direction in the model's activation space and project it out of the weights. No gradient steps, no training data. It is a weight edit, not a finetune. The assumption most people carry into…",
  "key_points": [
    "Abliterated models lose obedience before losing knowledge.",
    "Standard chat tests fail to detect hidden obedience degradation.",
    "Compliance rate metric measures format adherence independently of correctness."
  ],
  "editors_take": "Abliterated models' surprising loss of obedience before knowledge challenges the assumption of a tradeoff between compliance and intelligence, requiring new evaluation methods that prioritize format compliance and separate it from quality.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}