datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
claude-cowork-traces
Claude Cowork agent trace: AI stack regulation tweet + diagram
A single-session agent trace from Claude (Cowork mode), in raw Claude Code JSONL session format, natively supported by the Hugging Face agent trace viewer.
What happens in this session
The session covers an iterative content creation workflow:
Drafting a long tweet arguing that model weights, APIs, and apps sit at different layers of the AI stack and should be regulated differently
Generating a 16:9… See the full description on the dataset page: https://huggingface.co/datasets/clem/claude-cowork-traces.aicivs-npc-distillation
AICivs NPC distillation data
The data behind the AICivs student: teacher answers for the nine
operations of the AICivs service contract (a Minecraft mod whose villages are living civilizations), filtered the way the
game validates them, and formatted as the exact rows the student was trained on. Everything is synthetic: requests were
sampled from a 12-civilization world simulated headless for 120 seasons (souls, memories, chronicles, quest manifests),
answers were written by… See the full description on the dataset page: https://huggingface.co/datasets/cow9000/aicivs-npc-distillation.sd-webui-forge-configfreshair-corpus
Fresh Air interview corpus
NPR Fresh Air interview transcripts (2006-2026, ~1,070 episodes) plus derived
training/verification data, used by
whit3rabbit/freshair-analysis - an analysis
pipeline and a Claude Code "interviewer coach" skill trained on Terry Gross's measured
interviewing behavior.
License note
The pipeline code in the GitHub repo is 0BSD. This dataset is NPR-sourced transcript content
and derived annotations, not covered by that license - it's shared… See the full description on the dataset page: https://huggingface.co/datasets/cowWhySo/freshair-corpus.Hermes-Coworker-Flash
⚡ Hermes Coworker Flash – Fast, No‑Code Agent Conversations
Hermes Coworker Flash is a curated instruction‑style dataset built from the best “everyday assistant” traces of lambda/hermes-agent-reasoning-traces.It combines GLM‑5.1 and kimi-2.5 splits, keeping only the non‑programming, quick‑turnaround co‑worker tasks — the ones a user would ask a fast AI assistant, not a full‑fledged software engineer.
🧹 All chain‑of‑thought (<think>…</think>) has been removed so the model learns to… See the full description on the dataset page: https://huggingface.co/datasets/LiteMind/Hermes-Coworker-Flash.cowboy
🤠 Cowboy Voice Dataset
A conversational dataset designed to fine-tune language models to speak like a cowboy!
Each example contains a user question and a response written in authentic western slang,
with cowboy charm, frontier wisdom, and a whole lot of yeehaw!
This dataset was used to train the Voicebox Cowboy voice model:
https://huggingface.co/voice-box/cowboy
📊 Dataset Details
Property
Details
Size
300 examples
Format
JSONL
Language
English… See the full description on the dataset page: https://huggingface.co/datasets/quill-voice/cowboy.va-tool-calling-dataset
Voice Assistant Tool-Calling Dataset
Synthetic multi-turn conversations for fine-tuning a voice assistant with tool-calling capabilities.
Stats
Metric
Value
Total conversations
1585
Tool-call conversations
898
Direct-answer conversations
687
Tools defined
6
Generator model
anthropic/claude-haiku-4.5
Tools
Tool
Description
get_weather
Current weather and forecast
set_timer
Countdown timers and alarms… See the full description on the dataset page: https://huggingface.co/datasets/cowWhySo/va-tool-calling-dataset.CO-writer-127KDans-Prosemaxx-Cowriter-3-XLcoway_faqDans-Prosemaxx-Cowriter-3-LLiya_Reminder_Dataset_CorrectedDans-Prosemaxx-Cowriter-3-XSDans-Prosemaxx-Cowriter-3-SDans-Prosemaxx-Cowriter-3-MPocketDoc_Dans-Prosemaxx-Cowriter-XL-8192-shrunk-l3-qwq-all-aphrodite-filt