AI Tools.

Search
text generation by huihui-ai

Qwen2.5-72B-Instruct-abliterated

An abliterated (safety-removed) version of Qwen2.5-72B-Instruct by huihui-ai, where refusal mechanisms have been removed using directional activation manipulation. This allows the model to respond to requests the original would decline. The abliteration technique is reversible but the resulting model lacks safety guardrails.

Summary text generated by an automated pipeline from the model card · Not individually reviewed or run by us · How this page is made

Share: X Reddit HN LinkedIn

From the model card

Fields below are copied from the tags and counters on the HuggingFace repository huihui-ai/Qwen2.5-72B-Instruct-abliterated at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.

Publisher (HF namespace)
huihui-ai
Pipeline tag
text-generation
Library
Transformers
Weight formats
safetensors
License tag
other — read the license file in the repo before relying on it
Lineage
Language tags
Chinese (zho), English (eng), French (fra), Spanish (spa), Portuguese (por), German (deu), Italian (ita), Russian (rus), Japanese (jpn), Korean (kor), Vietnamese (vie), Thai (tha), Arabic (ara)
Downloads (HF counter at last fetch)
392,488
Likes (HF counter at last fetch)
49
Model card
https://huggingface.co/huihui-ai/Qwen2.5-72B-Instruct-abliterated

Use cases

  • Uncensored text generation for adult content platforms
  • Red-teaming and adversarial research in controlled settings
  • Creative writing without content policy restrictions
  • Research into LLM safety mechanism localization

Pros

  • 72B base model provides strong capability even after abliteration
  • multilingual support retained from Qwen2.5-72B
  • Transformers and TGI compatible
  • License carried over from base model

Cons

  • All safety refusals removed — actively generates harmful content without restriction
  • Abliteration can slightly degrade general reasoning quality
  • Requires same hardware as full 72B model — no compute saving
  • Not suitable for any deployment accessible to general users

Tags

transformerssafetensorsqwen2text-generationchatabliterateduncensoredconversationalzhoengfraspapordeuitarusjpnkorvietha