11 35 58

Agustín Piqueres Lajarín

plaguss

plaguss

AI & ML interests

None yet

Recent Activity

reacted to lewtun's post with 🔥 4 days 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 6 days ago

data-agents/jupyter-agent

liked a Space 17 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, 2024

• 32

Organizations

plaguss's activity

upvoted a paper 23 days ago

ProcessBench: Identifying Process Errors in Mathematical Reasoning

Paper • 2412.06559 • Published 25 days ago • 72

upvoted a paper 24 days ago

Free Process Rewards without Process Labels

Paper • 2412.01981 • Published Dec 2, 2024 • 29

upvoted a paper about 1 month ago

Solving math word problems with process- and outcome-based feedback

Paper • 2211.14275 • Published Nov 25, 2022 • 7

upvoted a collection about 1 month ago

SmolVLM

Collection

State-of-the-art compact VLMs for on-device applications: Base, Synthetic, and Instruct • 5 items • Updated 12 days ago • 30

upvoted 2 articles about 1 month ago

Article

Introducing Observers: AI Observability with Hugging Face datasets through a lightweight SDK

•

Nov 21, 2024

• 35

Article

Halo: Open Source Health Tracking with Wearables

•

Nov 19, 2024

• 99

upvoted a paper about 2 months ago

Aligning Large Language Models via Self-Steering Optimization

Paper • 2410.17131 • Published Oct 22, 2024 • 21

upvoted an article about 2 months ago

Article

Releasing the largest multilingual open pretraining dataset

•

Nov 13, 2024

• 98

upvoted an article 2 months ago

Article

Releasing Outlines-core 0.1.0: structured generation in Rust and Python

Oct 22, 2024

• 44

upvoted 2 articles 3 months ago

Article

How to build a custom text classifier without days of human labeling

•

Oct 17, 2024

• 55

Article

How to optimize your data labelling project with custom interfaces

•

Oct 16, 2024

• 18

upvoted a paper 3 months ago

DSBench: How Far Are Data Science Agents to Becoming Data Science Experts?

Paper • 2409.07703 • Published Sep 12, 2024 • 66

upvoted an article 4 months ago

Article

Fine-tuning a token classification model for legal data using Argilla and AutoTrain

•

Sep 7, 2024

• 14

upvoted 2 papers 4 months ago

Let Me Speak Freely? A Study on the Impact of Format Restrictions on Performance of Large Language Models

Paper • 2408.02442 • Published Aug 5, 2024 • 21

Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment

Paper • 2408.06266 • Published Aug 12, 2024 • 9

upvoted 4 articles 5 months ago

Article

Deploy Meta Llama 3.1 405B on Google Cloud Vertex AI

Aug 19, 2024

• 18

Article

Tool Use, Unified

Aug 12, 2024

• 70

Article

Introduction to ggml

Aug 13, 2024

• 124

Article

🔥 Argilla 2.0: the data-centric tool for AI makers 🤗

•

Jul 30, 2024

• 37

upvoted a collection 5 months ago

Llama 3.1 GPTQ, AWQ, and BNB Quants

Collection

Optimised Quants for high-throughput deployments! Compatible with Transformers, TGI & VLLM 🤗 • 9 items • Updated Sep 26, 2024 • 56