Abliteration releases a GLM-5.2 variant tuned for offensive cyber and agent testing
Effective_Attempt_72 · reddit · 2026-07-26
Released an abliterated version of GLM-5.2, fine-tuned for long adversarial and agent-testing trajectories.
- Reported evals: 97.5% benign utility on AgentDojo, 34.29% utility under attack, and 57.86% targeted ASR.
- The authors say coding ability stayed intact, with 81.2% on SWE-bench Verified and 80.1% on Terminal-Bench 2.1.
- The API is OpenAI/Anthropic-compatible, with zero retention by default and no baked-in refusals.
- They position it as useful for offensive cyber, red teaming, and testing agents that otherwise get refusal or drift mid-trajectory.
Related event: Abliterated GLM-5.2 Variant Released for Red Teaming(2 posts)→
More from Models
- Fern architecture claims 10M-token context with predictions still improving — AccBalanced · 2026-10-27
- Anthropic's follow-up on Claude's unauthorized web actions: self-policing fails — AccBalanced · 2026-10-27
- Local benchmark: Flint matches Qwen3.6 quality at 50/52 while running 3.6x faster on M4 Pro — RichardSocher · 2026-10-27
- ICML 2026 position paper: LLM 'thinking traces' are a harmful anthropomorphism — rao2z · 2026-10-27
- One stray 'u' and Claude starts spelling like a Brit — cantrell · 2026-10-27
- Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27