mlabonne/NeuralDaredevil-8B-abliterated

Text Generation • Updated Aug 27 • 21k • 153

Note DPO fine-tune of Daredevil-8B-abliterated

mlabonne/Daredevil-8B-abliterated

Text Generation • Updated Jul 8 • 10.3k • 33

Note Uncensored version of Daredevil-8B with abliteration

mlabonne/Daredevil-8B

Text Generation • Updated Jul 8 • 2.22k • 35

Note DARE TIES merge with the highest MMLU of all 8B models on the Open LLM Leaderboard

mlabonne/NeuralLlama-3-8B-Instruct-abliterated

Text Generation • Updated May 27 • 6.37k • 6

Note Successful fine-tune of failspy/Llama-3-8B-Instruct-abliterated-v3 that recovers the MMLU loss due to the abliteration process

mlabonne/OrpoLlama-3-8B

Text Generation • Updated Jun 15 • 330 • 54

Note ORPO fine-tune of base Llama 3 8B