MoRA: High-Rank Updating for Parameter-Efficient Fine-Tuning

Agentic-RAG explores advanced Retrieval-Augmented Generation systems enhanced with AI LLM agents.

Educational framework exploring ergonomic, lightweight multi-agent orchestration. Managed by OpenAI Solution team.

Curated list of datasets and tools for post-training.

This is the repository of HaluEval, a large-scale hallucination evaluation benchmark for Large Language Models.

Code and example data for the paper: Rule Based Rewards for Language Model Safety

This repo contains the source code for RULER: What’s the Real Context Size of Your Long-Context Language Models?

OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

Doing simple retrieval from LLM models at various context lengths to measure accuracy

Performs benchmarking on two Korean datasets with minimal time and effort.

Recipes to train reward model for RLHF.

Conifer: Improving Complex Constrained Instruction-Following Ability of Large Language Models

GPU Runner is a Python package that provides a simple way to run a function on a GPU.

Robust recipes to align language models with human and AI preferences

Distilabel is a framework for synthetic data and AI feedback for engineers who need fast, reliable and scalable pipelines based on verified research papers.

Notebooks and various random fun

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

Benchmarking LLMs with Challenging Tasks from Real Users

LLM101n: Let's build a Storyteller

Code and data for "KoDialogBench: Evaluating Conversational Understanding of Language Models with Korean Dialogue Benchmark" (LREC-COLING 2024)

Go ahead and axolotl questions

수능 국어 1등급에 도전하는 AI

日本語LLMまとめ - Overview of Japanese LLMs

Code base for internal reward models and PPO training

Our data munging code.

evolve llm training instruction, from english instruction to any language.

800,000 step-level correctness labels on LLM solutions to MATH problems

A collection of modular datasets generated by GPT-4, General-Instruct - Roleplay-Instruct - Code-Instruct - and Toolformer

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.