five

mats-10-sprint-cs-jb/loracles-fineweb-multidoc-qa

收藏
Hugging Face2026-04-18 更新2026-04-26 收录
下载链接:
https://hf-mirror.com/datasets/mats-10-sprint-cs-jb/loracles-fineweb-multidoc-qa
下载链接
链接失效反馈
官方服务:
资源简介:
--- pretty_name: Loracle FineWeb tags: - synthetic - fineweb - interpretability - question-answering - openrouter --- # loracles-fineweb-multidoc-qa Synthetic Loracle supervision data generated from FineWeb. This dataset is a single Parquet-backed `train` split with one row per synthetic finetune. This upload is a partial snapshot of a larger run. ## Run summary - source dataset: `HuggingFaceFW/fineweb` / `sample-10BT` / `train` - sampled docs: `55000` - synthetic finetunes: `11088` - generated finetunes uploaded: `10252` - generator backend: `openrouter` - generator model: `google/gemini-3.1-flash-lite-preview` - max docs per finetune: `20` - max token budget per finetune: `10000` - questions per finetune: `10` ## Table schema - one row per synthetic finetune - `dpo_behavior`: one PKU-SafeRLHF-style harm category or `None` - nested `questions` field with `10` ordered question-answer pairs - finetune metadata merged in from `synthetic_models.jsonl` - stored on the Hub as Parquet ## Local artifacts - run directory: `/ceph/scratch/jbauer/loracle/fineweb_openrouter_flash31lite_topics10kbase_v1_20260417` - sample manifest: `sample_manifest.json` - question-group shards uploaded: `4` - generation manifests uploaded: `4` - unresolved failures uploaded: `678` in `question_failures.final.jsonl` - auxiliary files kept as repo artifacts: `synthetic_models.jsonl`, `sampled_docs.jsonl`

### 数据集展示名:洛拉科莱FineWeb(Loracle FineWeb) #### 标签: - 合成数据 - FineWeb - 可解释性 - 问答 - OpenRouter --- # loracles-fineweb-multidoc-qa 本数据集为基于FineWeb生成的合成洛拉科莱监督数据。 本数据集仅包含一个基于Parquet格式存储的训练拆分(train split),每条数据对应一个合成微调任务。本次上传仅为完整运行流程的部分快照。 ## 运行概况 - 源数据集:`HuggingFaceFW/fineweb` / `sample-10BT` / `train` - 采样文档数:55000 - 合成微调任务总数:11088 - 已上传生成的微调任务数:10252 - 生成后端:OpenRouter - 生成模型:`google/gemini-3.1-flash-lite-preview` - 单微调任务最大文档数:20 - 单微调任务最大Token(Token)预算:10000 - 单微调任务问答对数量:10 ## 数据表结构 - 每条数据对应一个合成微调任务 - `dpo_behavior`:PKU-SafeRLHF风格的伤害类别标签或`None` - 嵌套的`questions`字段,包含10组有序问答对 - 微调元数据从`synthetic_models.jsonl`中合并获取 - 以Parquet格式存储于Hugging Face Hub ## 本地工件 - 运行目录:`/ceph/scratch/jbauer/loracle/fineweb_openrouter_flash31lite_topics10kbase_v1_20260417` - 采样清单文件:`sample_manifest.json` - 已上传的问答组分片数:4 - 已上传的生成清单数:4 - 未解决的失败任务共678条,存储于`question_failures.final.jsonl` - 作为仓库工件留存的辅助文件:`synthetic_models.jsonl`、`sampled_docs.jsonl`
提供机构:
mats-10-sprint-cs-jb
二维码
社区交流群
二维码
科研交流群
商业服务