Uncensored LFM3 fork drops refusal rate from 61.7% to 0% in a 1.1GB file
alexcovo_eth · x · 2026-11-13
A community fork of Liquid AI's 1.7B chat model LFM3 strips out safety guardrails, with the creator claiming refusals fell from 61.7% to 0% on their test set. The GGUF release uses the LFM open license with no gatekeeping, has hit 132k downloads, and the Q4 quant is just 1.1GB — small enough to run locally on nearly any PC.
More from Models
- Anthropic's internal model 4 runs ~6 weeks ahead of AI-2027 curve, claim circulates — haider1 · 2026-11-13
- Edge0-145B: an open-source 145B-parameter model claiming frontier capabilities on your own computer — BrightCandle · 2026-11-13
- Inkling 2 Small architecture detailed: attention residuals, NVFP4/MXFP8 mixed-precision experts — teortaxesTex · 2026-11-13
- Inkling 2 Small scores 31 on AA index, tops GLM and DeepSeek on finance evals — simonguozirui · 2026-11-13
- Gemini glitch leaks its backend response instructions to the user in plain text — SushiManPlatinum · 2026-11-13
- InferenceBench: agents beat vLLM defaults but lose to simple hyperparameter search — maksym_andr · 2026-11-13