kmseong/llama3.2_3b_SSFT_epoch5_lr5e-5
TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kPublished:Apr 4, 2026License:llama3.2Architecture:Transformer Loading

The kmseong/llama3.2_3b_SSFT_epoch5_lr5e-5 is a 3.2 billion parameter Llama 3.2-based causal language model developed by kmseong. This model has undergone Phase 0 of Safety-WaRP (Weight space Rotation Process) using the Circuit Breakers dataset, focusing on base safety training. It is specifically fine-tuned to establish safety mechanisms and generate refusal responses to harmful prompts. While optimized for safety, its utility for general tasks like reasoning or mathematics may be reduced at this stage.

Loading preview...