11 34 58

Agustín Piqueres Lajarín

plaguss

plaguss

AI & ML interests

None yet

Recent Activity

reacted to lewtun's post with 🔥 1 day ago

This paper (https://huggingface.co/papers/2412.18925) has a really interesting recipe for inducing o1-like behaviour in Llama models: * Iteratively sample CoTs from the model, using a mix of different search strategies. This gives you something like Stream of Search via prompting. * Verify correctness of each CoT using GPT-4o (needed because exact match doesn't work well in medicine where there are lots of aliases) * Use GPT-4o to reformat the concatenated CoTs into a single stream that includes smooth transitions like "hmm, wait" etc that one sees in o1 * Use the resulting data for SFT & RL * Use sparse rewards from GPT-4o to guide RL training. They find RL gives an average ~3 point boost across medical benchmarks and SFT on this data already gives a strong improvement. Applying this strategy to other domains could be quite promising, provided the training data can be formulated with verifiable problems!

liked a Space 3 days ago

data-agents/jupyter-agent

liked a Space 15 days ago

HuggingFaceH4/blogpost-scaling-test-time-compute

View all activity

Articles

How we leveraged distilabel to create an Argilla 2.0 Chatbot

Jul 16

• 32

Organizations

plaguss's activity

liked a Space 3 days ago

Running

144

🏃

Jupyter Agent

liked a Space 15 days ago

Running

408

📈

Scaling test-time compute

liked a model 21 days ago

meta-llama/Llama-3.1-8B

Text Generation • Updated Oct 16 • 664k • 1.23k

liked a dataset 22 days ago

HuggingFaceFW/fineweb-2

Viewer • Updated 23 days ago • 13.8B • 97k • 371

liked a dataset about 1 month ago

data-is-better-together/open-image-preferences-v1-results

Viewer • Updated 22 days ago • 10k • 142 • 27

liked a model about 1 month ago

HuggingFaceTB/SmolVLM-Instruct

Image-Text-to-Text • Updated 29 days ago • 70.3k • 303

liked a Space about 1 month ago

Running on Zero

105

📊

SmolVLM

liked a dataset about 1 month ago

AI-MO/NuminaMath-CoT

Viewer • Updated Nov 25 • 860k • 4.68k • 282

liked a Space about 1 month ago

Running

🖼️

Image Preferences - Argilla annotation space

A community project to create an image preferences dataset.

liked 2 datasets about 1 month ago

HuggingFaceTB/smol-smoltalk

Viewer • Updated Nov 21 • 485k • 932 • 22

HuggingFaceTB/smoltalk

Viewer • Updated Nov 26 • 2.2M • 10.8k • 261

liked 2 models about 2 months ago

HuggingFaceTB/SmolLM2-1.7B-Instruct

Text Generation • Updated 27 days ago • 96.1k • • 449

HuggingFaceTB/SmolLM2-1.7B

Text Generation • Updated Nov 24 • 12.3k • 85

liked a dataset 2 months ago

neuralwork/arxiver

Viewer • Updated Nov 1 • 63.4k • 525 • 355

liked 6 datasets 3 months ago