中文简介
该仓库保存 Claude Fable 5 相关的匿名化原始智能体轨迹,并说明其与 Glint-Research/Fable-5-traces 存在相同数据来源,二者不应重复混用。上游建议使用专用工具转换和过滤会话;当前许可证信息不完整,且仍需检查匿名化、工具输出与第三方内容。
上游模型卡 / 数据集卡
该仓库保存 Claude Fable 5 相关的匿名化原始智能体轨迹,并说明其与 Glint-Research/Fable-5-traces 存在相同数据来源,二者不应重复混用。上游建议使用专用工具转换和过滤会话;当前许可证信息不完整,且仍需检查匿名化、工具输出与第三方内容。
claude-fable-5 Agent Traces It's worth noting that our team was working with **Glint-Research** to collect as much fable data as possible. These are just the anonymized raw traces of both of our teams combined. This means that Glint-Research/Fable-5-traces was created from formatting and splitting up this same dataset. If you use one for your tune, don't use the other (it's the same exact data). For training on this dataset I recommend using the teich package to convert to openai style chats. It parses and filters out things like hitting limits and model switches, etc. As well as knows the exact tool-schemas and descriptions for the tools used. I encourage everyone to please upload their fable-5 traces from whatever harness they may have used. This is all valuable data and we need to pool it together to make something meaningful :) To easily extract and anonymize your data you can use teich as well: Check out teich on github for more info about all that the package can do! https://github.com/TeichAI/teich Tool schema snapshot Training-ready tool schema snapshot
上游文件元数据
.gitattributes2.52 KB004c0d63-c96f-431d-8cf8-f78ca9367eb5.jsonl2.09 MB06ec42c3-2184-40c5-b0ee-98c3235b4c4c.jsonl435.77 KB17643245-7004-426a-b3d0-40a5fb6fd397.jsonl1.94 MB189f0549-c2fc-4fc7-b50d-37cd255eabf2.jsonl1.33 MB1afc3cc3-792e-49a2-83c4-19ddb66ba008.jsonl1.20 MB1d29fdca-8b1a-4087-b14f-7c4e69701fbe.jsonl168.29 KB21576682-1c94-4e42-a9d2-0f9e0137ffa7.jsonl44.96 KB2f6b46c2-a83e-43e7-9012-76c7b434abba.jsonl806.27 KB311883ad-f031-4619-ab72-1aeb410a9123.jsonl23.10 KB32bd3eac-d079-4d7f-8f49-b34180656abd.jsonl918.24 KB32c96eea-ab6f-4268-9f48-60fb2c8838e7.jsonl1.48 MB
本页面为橙子AI科技的中文整理与服务说明,不代表资源作者或平台官方页面。实际许可、访问和使用条件以上游原文为准。