Nathan's picture

Open to Collab

Nathan

DreamFast

·

https://nathan.sapwell.net/

AI & ML interests

Local LLM

Recent Activity

updated a model about 7 hours ago

DreamFast/Qwen3-VL-8B-Heretic-1.3.0

published a model about 8 hours ago

DreamFast/Qwen3-VL-8B-Heretic-1.3.0

updated a collection 2 days ago

View all activity

Organizations

New activity in DuoNeural/Gemma-4-E2B-Heretic 5 days ago

Benchmarking and KL Divergence Difference

#1 opened 5 days ago by

New activity in llmfan46/gemma-4-E2B-it-ultra-uncensored-heretic 5 days ago

Independent verification: lowest KL of all 13 variants, capability fully preserved

#1 opened 5 days ago by

New activity in TrevorJS/gemma-4-E2B-it-uncensored 5 days ago

Independent verification: 99.5% ASR with capability preserved

#2 opened 5 days ago by

New activity in coder3101/gemma-4-E4B-it-heretic 5 days ago

Independent benchmark verification: your Gemma 4 variant is the top pick

#2 opened 5 days ago by

New activity in WWTCyberLab/gemma-4-E2B-it-abliterated 5 days ago

Independent verification of refusal rate and language modelling quality

#2 opened 5 days ago by

New activity in wangzhang/gemma-4-E2B-it-abliterated 5 days ago

Analysis of wangzhang/gemma-4-E2B-it-abliterated compared to others

#2 opened 5 days ago by

New activity in treadon/gemma4-E2B-it-Abliterated-AND-Disinhibited-USE-THIS 5 days ago

Independent verification of KL divergence and capability impact

#1 opened 5 days ago by

New activity in DreamFast/Qwen3.6-27B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark 8 days ago

Basically base on that analysis, Huihui is better than ton of finetunes in throughput

#4 opened 9 days ago by

New activity in DreamFast/Qwen3.6-27B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark 11 days ago

NVFP4 version?

#3 opened 15 days ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16 13 days ago

Benchmarks and Chain of Thought

#8 opened 19 days ago by

New activity in wangzhang/Qwen3.6-27B-abliterated 19 days ago

Abliterix and other abliteration techniques

#1 opened 19 days ago by

New activity in DreamFast/Qwen3.6-27B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark 22 days ago

Thank you

#1 opened 22 days ago by

New activity in HauhauCS/Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive about 1 month ago

Would you provide a safetensor version of models?

#8 opened 3 months ago by

New activity in HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive about 1 month ago

safetensor version

#10 opened about 2 months ago by

"These are meant to be the best lossless uncensored models out there."

#3 opened about 2 months ago by

New activity in HauhauCS/Qwen3.6-27B-Uncensored-HauhauCS-Aggressive about 1 month ago

safetensor?

#4 opened about 1 month ago by

New activity in DreamFast/gemma-3-12b-it-heretic-v2 about 1 month ago

The model won't generate text if image input is enabled with TextGenerateLTX2Prompt

#1 opened about 1 month ago by

New activity in DontPlanToEnd/UGI-Leaderboard about 2 months ago

Eval request: HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive

#590 opened 3 months ago by

New activity in DreamFast/gemma-3-12b-it-heretic 3 months ago

why vision tower layers not included in comfyui version?

#5 opened 3 months ago by

NVFP4 is supported now please provide one.

#4 opened 3 months ago by