💬 Community: Join the Abliterlitics Discord for discussion, model releases and support. > Forensic analysis by Abliterlitics — open-source abliteration forensics toolkit This is the HauhauCS abliteration of GLM-4.7-Flash, converted from the BF16 GGUF release to native safetensors using ungguf. HauhauCS claims these are «No changes to datasets or capabilities. Fully functional, 100% of what the original authors intended, just without the refusals» and describes them as «the best lossless uncensored models out there.» I ran the full forensic suite to find out. Benchmarks, safety evaluation, weight analysis, the works. And I compared against three other abliteration techniques applied to the same base model: Heretic (by trohrbaugh, using the Heretic tool by p-e-w), Huihui, and Abliterix. Abliterix is a custom variant built on Heretic with router and shared expert targeting. GLM-4.7-Flash is a Mixture-of-Experts reasoning model with 64 routed experts per layer plus shared experts. It uses Multi-head Latent Attention rather than standard multi-head attention. The model has a reasoningcontent` field for chain-of-thought, making it a thinking/reasoning model. This has significant…
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: DreamFast
Теги: glm4_moe_lite, uncensored, abliterated, glm4, moe, conversational, en, zh
Лайков: 3 | Загрузок: 48
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
