7 11 219

Yichao 'Peak' Ji

peakji

AI & ML interests

Agents, Small Language Models, Retrieval-Augmented Generation, Information Extraction

Recent Activity

liked a model 15 days ago

Qwen/Qwen2.5-VL-32B-Instruct

liked a dataset 15 days ago

starvector/text2svg-stack

liked a model 19 days ago

mistralai/Mistral-Small-3.1-24B-Instruct-2503

View all activity

Organizations

peakji's activity

liked a model 15 days ago

Qwen/Qwen2.5-VL-32B-Instruct

Image-Text-to-Text • Updated 3 days ago • 314k • 323

liked a dataset 15 days ago

starvector/text2svg-stack

Viewer • Updated Jan 10 • 2.18M • 816 • 8

liked 2 models 19 days ago

mistralai/Mistral-Small-3.1-24B-Instruct-2503

Image-Text-to-Text • Updated 1 day ago • 125k • 1.08k

google/gemma-3-27b-it

Image-Text-to-Text • Updated 19 days ago • 1.03M • • 1.13k

reacted to clem's post with ❤️ about 1 month ago

Post

7284

I was chatting with @peakji , one of the cofounders of Manu AI, who told me he was on Hugging Face (very cool!).

He shared an interesting insight which is that agentic capabilities might be more of an alignment problem rather than a foundational capability issue. Similar to the difference between GPT-3 and InstructGPT, some open-source foundation models are simply trained to 'answer everything in one response regardless of the complexity of the question' - after all, that's the user preference in chatbot use cases. Just a bit of post-training on agentic trajectories can make an immediate and dramatic difference.

As a thank you to the community, he shared 100 invite code first-come first serve, just use “HUGGINGFACE” to get access!