abliteration

English

Etymology

Blend of ablate +‎ obliteration, see abliterate.

Noun

abliteration (countable and uncountable, plural abliterations)

  1. (neologism, artificial intelligence, computing) The process of uncensoring a large language model by modifying internal functions to eliminate refusal behaviors while preserving the remaining functions of the model.
    • 2024 June 13, Maxime Labonne, “Uncensor any LLM with abliteration”, in Hugging Face Blog‎[1], retrieved 29 May 2025:
      We applied abliteration to Daredevil-8B to uncensor it, which also degraded the model's performance.
This article is issued from Wiktionary. The text is available under Creative Commons Attribution-Share Alike 4.0 unless otherwise noted. Additional terms may apply for the media files.