Unverified50% confidenceReleaseExact time
This work is fully open-sourced, including the capture proxy in OpenEnv, the trainer in TRL, the task set, SFT data, training code, and all seven trained models.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/6/2026
First Seen
Valid until: 1/4/2027
Sources
Related Entities
Related Claims
UnverifiedEleutherAI的GPT-NeoX公开了训练代码、数据集The Pile和训练日志等全部要素，属于完全开源75% similarUnverifiedFull open source requires the public disclosure of training datasets, training code, evaluation frameworks, and all technical documentation, theoretically allowing anyone to reproduce the entire model from scratch.72% similarUnverifiedOpenAI和Anthropic的大模型都大量使用了GitHub开源代码、Stack Overflow问答、技术文档等公开数据进行训练71% similarUnverifiedUnlike closed-source API services, open-weight models allow developers to have full control over the entire process of model training, fine-tuning, and deployment.71% similarVerified开放权重模型仅发布训练完成的模型参数文件，而严格意义的开源AI要求公开训练代码、数据集与完整复现流程70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/980177API
curl https://kongchang.com/api/v1/knowledge/claims/980177MCP
get_claim(id=980177)