CValues
Model value alignment
Evaluates and aligns the values of Chinese large language models with safety and responsibility standards
面向中文大模型价值观的评估与对齐研究
481 stars
1 watching
20 forks
Language: Python
last commit: over 1 year ago benchmarkchinese-llmsevaluationhuman-valuesllmsmulti-choiceresponsibilitysafety
Related projects:
| Repository | Description | Stars |
|---|---|---|
 ys-zong/vlguard | Improves safety and helpfulness of large language models by fine-tuning them using safety-critical tasks | 47 |
 rlhf-v/rlhf-v | Aligns large language models' behavior through fine-grained correctional human feedback to improve trustworthiness and accuracy. | 245 |
 felixgithub2017/mmcu | Measures the understanding of massive multitask Chinese datasets using large language models | 87 |
 pku-alignment/align-anything | Aligns large multimodal models with human intentions and values using various algorithms and fine-tuning methods. | 270 |
 applieddatasciencepartners/xgboostexplainer | Provides tools to understand and interpret the decisions made by XGBoost models in machine learning | 253 |
 brightmart/xlnet_zh | Trains a large Chinese language model on massive data and provides a pre-trained model for downstream tasks | 230 |
 ailab-cvc/seed-bench | A benchmark for evaluating large language models' ability to process multimodal input | 322 |
 ethicalml/xai | An eXplainability toolbox for machine learning that enables data analysis and model evaluation to mitigate biases and improve performance | 1,135 |
 ethanyanjiali/minchatgpt | This project demonstrates the effectiveness of reinforcement learning from human feedback (RLHF) in improving small language models like GPT-2. | 214 |
 princeton-nlp/charxiv | An evaluation suite for assessing chart understanding in multimodal large language models. | 85 |
| x-plug/mplug-halowl | Evaluates and mitigates hallucinations in multimodal large language models | 82 |
 cbcrg/benchfam | Generates a benchmark dataset for evaluating protein alignment programs | 3 |
 aidc-ai/ovis | An MLLM architecture designed to align visual and textual embeddings through structural alignment | 575 |
 yuweihao/mm-vet | Evaluates the capabilities of large multimodal models using a set of diverse tasks and metrics | 274 |
 hit-scir/chinese-mixtral-8x7b | An implementation of a large language model for Chinese text processing, focusing on MoE (Multi-Headed Attention) architecture and incorporating a vast vocabulary. | 645 |