www.experientiallabs.ai
网页
综合
Cyber AI: Uncensored Models for Security Work
…ion in the weights that produces refusals. An abliterated GLM-5.3 Flash published by dealignai. It scores 87.3 on MMLU against 86.7 for the base. An abliterated DeepSeek V4.1 Flash published…
2026-11-14
自动蜘蛛
相关度 11.6