An Easy-to-use, Scalable and High-performance RLHF Framework (70B+ PPO Full Tuning & Iterative DPO & LoRA & RingAttention & RFT)

This is a replicate of DeepSeek-R1-Zero and DeepSeek-R1 training on small models with limited data

Clean, accessible reproduction of DeepSeek R1-Zero

Inspect: A framework for large language model evaluations

Training Sparse Autoencoders on Language Models

Entropy Based Sampling and Parallel CoT Decoding

Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends

smolLM with Entropix sampler on pytorch

Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering

A framework for few-shot evaluation of language models.

Experimental spaced repetition platform for exploring ideas in memory augmentation and programmable attention

smol-podcaster is your autonomous podcast production intern 🐣

SGLang is a fast serving framework for large language models and vision language models.

Optimizing inference proxy for LLMs

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Video+code lecture on building nanoGPT from scratch

AI Chat Browser: Fast, Full webapp access to ChatGPT / Claude / Bard / Bing / Llama2! I use this 20 times a day.

🐚 OpenDevin: Code Less, Make More

Distilabel is a framework for synthetic data and AI feedback for AI engineers that require high-quality outputs, full data ownership, and overall efficiency.

An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.

LlamaIndex is a data framework for your LLM applications

Hackable implementation of state-of-the-art open-source LLMs based on nanoGPT. Supports flash attention, 4-bit and 8-bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-l…

Reference implementation for DPO (Direct Preference Optimization)