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

Arena-Hard-Auto: An automatic LLM benchmark.

LMDeploy is a toolkit for compressing, deploying, and serving LLMs.

Is Your Model Really A Good Math Reasoner? Evaluating Mathematical Reasoning with Checklist

Implementation for "Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs"

整理开源的中文大语言模型，以规模较小、可私有化部署、训练成本较低的模型为主，包括底座模型，垂直领域微调及应用，数据集与教程等。

中文 NLP 预处理、解析工具包，准确、高效、易用 A Chinese NLP Preprocessing & Parsing Package www.jionlp.com

Robust recipes to align language models with human and AI preferences

[ACL2023] We introduce LLM-Blender, an innovative ensembling framework to attain consistently superior performance by leveraging the diverse strengths of multiple open-source LLMs. LLM-Blender cut …

An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.

Bob 是一款 macOS 平台的翻译和 OCR 软件。

Towards Real-World Writing Assistance: A Chinese Character Checking Benchmark with Faked and Misspelled Characters

The repository of EMNLP 2023 "MixEdit: Revisiting Data Augmentation and Beyond for Grammatical Error Correction"

[AAAI 2024] MESED: A Multi-modal Entity Set Expansion Dataset with Fine-grained Semantic Classes and Hard Negative Entities

Automatically split your PyTorch models on multiple GPUs for training & inference

Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

[ICLR 2021] Contrastive Learning with Adversarial Perturbations for Conditional Text Generation

A high-throughput and memory-efficient inference and serving engine for LLMs

Code for ACL 2024 paper "TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space"

主要记录大语言大模型（LLMs） 算法（应用）工程师相关的知识及面试题

A natural language interface for computers

Reference implementation for DPO (Direct Preference Optimization)

Dataset and evaluation script for "Evaluating Hallucinations in Chinese Large Language Models"

Knowledge Verification to Nip Hallucination in the Bud

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.

A framework for few-shot evaluation of language models.

A package to evaluate factuality of long-form generation. Original implementation of our EMNLP 2023 paper "FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation"