Hugging Face's logo
Alex's picture

Alex

Asystemoffields
TaiwanLZX's profile picture
Ā·

AI & ML interests

Compression, creatively reconceiving problems, RL, interpretability, model architecture, model modification, democratization of AI, machines of loving grace type scenarios.

Recent Activity

reacted to black-yt's post with šŸ”„ 1 day ago
Hey all ā€” our ResearchClawBench leaderboard just updated šŸ”„ We let AI do real science: 40 tasks across 10 disciplines, compared to human papers. Hard example? šŸ”ļø Glacier mass change ā€” AI must integrate 233 datasets from 35 teams, 4 methods, reproduce 6542Ā±387 Gt ice loss vs IPCC. No toy problems. Latest leaderboard (2026-06-09) šŸ“Š: Agents: šŸ„‡ Claude Code 21.5 (50 = match human), $5.3; šŸ„ˆ EvoScientist 18.8, $4.1; šŸ„‰ Codex CLI 18.4, just $2.0 LLMs+Harness: šŸ„‡ Claude-Opus-4.8 21.1, $4.0; šŸ„ˆ Claude-Opus-4.7 20.7; šŸ„‰ MiniMax-M3 19.8, only $0.45; Qwen3.7-Max 18.7, $0.42, 11min šŸ’„ Claude still king, but MiniMax/Qwen/DeepSeek are crazy cheap and competitive. Expensive isn't always better. šŸ“Ž Code & star: https://github.com/InternScience/ResearchClawBench šŸ  Website: https://internscience.github.io/ResearchClawBench-Home/ šŸ¤— Upvote paper: https://huggingface.co/papers/2606.07591
reacted to danielhanchen's post with šŸ”„ 5 days ago
Gemma 4 12B can now run locally on just 8GB RAM via Dynamic GGUFs. Google's new model, Gemma 4 12B Unified supports image, audio and 256K context. You can run and train the model via Unsloth Studio. GGUF: https://huggingface.co/unsloth/gemma-4-12b-it-GGUF Guide: https://unsloth.ai/docs/models/gemma-4
updated a model 8 days ago
Asystemoffields/gemma-4-E2B-it-PMRA-GGUF
View all activity

Organizations

None yet