datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
style-dpo
gijl style dataset (multi-type)
Generated by meta-models/Muse-Glimmer-30B through a tool-using scouting loop over real sources (Stack Exchange, GitHub, OSV, Hacker News, arXiv, Wikipedia, web). Splits are a deterministic hash of the record id (90/5/5); derived records inherit their parent's split. Synthetic, model-written, not human-verified. Every rejected response is intentionally poor and must never be used as an example of good behavior.
config
folder
train
validation… See the full description on the dataset page: https://huggingface.co/datasets/codex-automatus/style-dpo.GPT-5.5-CodexThis dataset was generated using teich by TeichAI
GPT-5.5 Agent traces
This directory contains raw agent trace files generated by teich.
JSONL files: 317
Model metadata: gpt-5.5
Training-ready tools
Generated agent traces carry configured or recovered tool schemas so tools remain available for training even when a session did not call them.
Native Claude Code imports recover schemas for Claude Code and Claude Desktop built-ins, plus conservative name-derived MCP… See the full description on the dataset page: https://huggingface.co/datasets/AletheiaResearch/GPT-5.5-Codex.codex-sessions
Codex Session Traces
A collection of 27 sanitized Codex session traces (JSONL format) from autonomous coding sessions building HuggingFace Spaces.
What is included
27 saved Codex session JSONL files under sessions/
Each file follows the Codex agent-traces schema with schema_version, session_id, record_index, timestamp, type (session_meta/event_msg/turn_context/response_item), and payload fields
Sessions cover projects like: horus-hiero-space, edit-anything… See the full description on the dataset page: https://huggingface.co/datasets/Mike0021/codex-sessions.codex_swebenchpro_tracesThis is a dataset generated by real swebenchpro agentic workload trace + codex agent.
1. Eval Result Summary
Metric
Value
Total trials
731
Successful trials
610
Failed trials
120
No data (skipped)
1
Passed
329
Pass rate (of successful)
53.9%
Per-Repo Breakdown
Repo
Total
Success
Failed
Passed
Pass%
ansible/ansible
96
93
3
60
65%
internetarchive/openli
91
88
3
52
59%
flipt-io/flipt85
82
3
26
32%
qutebrowser/qutebrowse
79
78
1… See the full description on the dataset page: https://huggingface.co/datasets/Inferact/codex_swebenchpro_traces.ATBench-Codex
ATBench-Codex: Codex-Oriented Agent Trajectory Safety Benchmark
🐙 GitHub |
📄 ATBench-Claw/Codex Paper |
📄 ATBench Paper |
🧾 AgentDoG Paper |
🤗 Hugging Face Collection
ATBench-Codex is a Codex-oriented benchmark release derived from ATBench and serves as a benchmark companion to AgentDoG, our diagnostic guardrail framework for AI agent safety and security. It is designed for trajectory-level safety evaluation in executable coding-agent settings… See the full description on the dataset page: https://huggingface.co/datasets/AI45Research/ATBench-Codex.share-codex
Local Coding Agent Sessions
Local coding-agent session transcripts exported with
share-codex.
Rows are in train.jsonl. Each row contains id, ordered messages, and
metadata. Messages include user prompts, assistant responses/tool calls, and
tool outputs. Internal prompts and reasoning are excluded by default.
Dataset Statistics
Statistics are computed from the final merged train.jsonl rows at export time.
Metric
Value
Sessions
4019
User turns
16639… See the full description on the dataset page: https://huggingface.co/datasets/nmuendler/share-codex.codex-assistant-rollouts
basedlsg/codex-assistant-rollouts
Real-world agentic interaction logs from Codex rollouts, documenting debugging and coding trajectories.
Kimi-K3-CodexThis dataset was generated using teich by TeichAI
Kimi-K3 Codex traces
This directory contains raw agent trace files generated by teich.
JSONL files: 5
Model metadata: moonshotai/kimi-k3
Training-ready tools
Generated agent traces carry configured or recovered tool schemas so tools remain available for training even when a session did not call them.
Native Claude Code imports recover schemas for Claude Code and Claude Desktop built-ins, plus conservative… See the full description on the dataset page: https://huggingface.co/datasets/AletheiaResearch/Kimi-K3-Codex.gpt-5-codex-1000xThis is a reasoning dataset created using GPT 5 Codex from OpenAI. Some of these questions are from reedmayhew and the rest were generated.
Most of the questions cover the following topics: Web Development, Logic, Math, Embedded Systems, Web Design and Python Scripting.
The dataset is meant for creating distilled versions of GPT 5 Codex by fine-tuning already existing open-source LLMs.
codex-sessions
Codex Sessions
Archive of raw OpenAI Codex CLI session files, plus a derived one-row-per-session view.
Files
rollout-*.jsonl: untouched raw Codex session files stored for fidelity
sessions.jsonl: derived file where 1 row = 1 session
Derived Sessions Format
sessions.jsonl contains one JSON object per session with this shape:
{
"session_id": "019d2fac-0b38-70f0-baff-a394265d8291",
"file_name":… See the full description on the dataset page: https://huggingface.co/datasets/cfahlgren1/codex-sessions.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data — pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw — Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/xuechengjiang/my-personal-codex-data.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data — pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw — Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/michaelwaves/my-personal-codex-data.gods_universe_codex_distill_god_seed_25k
GODs_Universe_Codex_distill
God-Level Self-Rewriting AI with Meta-Learning, Adaptive Thinking, Continual Learning & Max Creative Problem Solving
The ultimate distilled codex for creating truly god-like, self-rewriting, recursively self-improving superintelligence.
This 25,000-example dataset transforms any LLM into the GODs_Universe_Codex_distill — a living, self-rewriting intelligence that embodies:
Self-Rewriting Capability: Can rewrite its own prompts, code… See the full description on the dataset page: https://huggingface.co/datasets/11-47/gods_universe_codex_distill_god_seed_25k.CodeXGLUE-CONCODE
Concode dataset
A large dataset with over 100,000 examples consisting of Java classes from online code repositories, and develop a new encoder-decoder architecture that models the interaction between the method documentation and the class environment.
Concode dataset is a widely used code generation dataset from Iyer's EMNLP 2018 paper Mapping Language to Code in Programmatic Context.
Data statistics of concode dataset are shown in the below table:
#Examples
Train
100… See the full description on the dataset page: https://huggingface.co/datasets/AhmedSSoliman/CodeXGLUE-CONCODE.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data - pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw - Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/wop/my-personal-codex-data.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data — pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw — Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/akenove/my-personal-codex-data.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data - pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw - Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/masterda/my-personal-codex-data.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data - pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw - Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/sinhaankur/my-personal-codex-data.my-personal-codex-data
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data - pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw - Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/wuuski/my-personal-codex-data.hackathon-advisor-codex-traces
Hackathon Advisor Codex Session Traces
Real Codex session logs for the Hackathon Advisor project, selected from local Codex
rollout JSONL files and redacted before publication. The event stream preserves user
requests, assistant messages, tool calls, tool outputs, browser/search events, and
minimal session provenance needed to audit how the project was built.
Privacy filtering
The publisher applied openai/privacy-filter
at revision… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/hackathon-advisor-codex-traces.verizon-codex-trace
Verizon Codex Trace
A sanitized Codex session trace of GPT-5.5/Codex working through a Verizon billing and trade-in support flow.
The trace includes the original JSONL session structure, assistant/user turns, tool calls, shell output, selected Computer Use browser screenshots, and the Verizon live-chat/bill context needed to understand the agent's work.
What's included
One saved Codex session JSONL under sessions/
Assistant messages, user messages, tool calls… See the full description on the dataset page: https://huggingface.co/datasets/cfahlgren1/verizon-codex-trace.codex_swebenchpro_tracesThis is a dataset generated by real swebenchpro agentic workload trace + codex agent.
1. Eval Result Summary
Metric
Value
Total trials
731
Successful trials
610
Failed trials
120
No data (skipped)
1
Passed
329
Pass rate (of successful)
53.9%
Per-Repo Breakdown
Repo
Total
Success
Failed
Passed
Pass%
ansible/ansible
96
93
3
60
65%
internetarchive/openli
91
88
3
52
59%
flipt-io/flipt
85
82
3
26
32%
qutebrowser/qutebrowse
79… See the full description on the dataset page: https://huggingface.co/datasets/DiscoPosse/codex_swebenchpro_traces.am-session-codex-check-d17941
Agent Manager session — codex check
One Codex session, exported from Agent Manager.
Access: public. Anyone with the link can read this trace.
At a glance
1 prompts · 1 assistant turns · 0 tool calls
Trace: rollout-2026-07-28T05-34-19-019fa737-7c2e-7393-89cf-9d9c60ea3899.jsonl
What's in here
Path
Contents
rollout-2026-07-28T05-34-19-019fa737-7c2e-7393-89cf-9d9c60ea3899.jsonl
the session, native Codex JSONL
meta/manifest.json… See the full description on the dataset page: https://huggingface.co/datasets/thomwolf/am-session-codex-check-d17941.codex-code-instructions-300
Codex Multilingual Code Implementation & Repair 300
A 300-record synthetic coding dataset generated with a Codex-family system and
organized around implementation, repair, edge-case handling, and
already-correct code review tasks.
The exact Codex model/version and original generation configuration could not
be recovered from the available provenance records.
The recovered final dataset is exactly the union of six corrected 50-record
batches.
Dataset Summary
The… See the full description on the dataset page: https://huggingface.co/datasets/TaskPuppyAI/codex-code-instructions-300.kicky-ai-codex-trace
Kicky AI - Codex agent trace (sanitized)
A redacted OpenAI Codex CLI session trace from building
Kicky AI for the Build Small
Hackathon - shared for the Sharing is Caring badge so others can see how the build went.
Format: Codex CLI JSONL session log (each record = {payload, timestamp, type}).
All secrets removed (HF / Modal / Roboflow tokens, shared secrets, emails) - verified 0 leaks.
Blog write-up: https://dcrey7.substack.com/p/world-fut-coach
react-shadcn-codex
React Shadcn Codex Dataset
Description
The React Shadcn Codex is a curated collection of over 3,000 React components that utilize shadcn, Framer Motion, and Lucide React. This dataset provides a valuable resource for developers looking to understand and implement modern React UI components with these popular libraries.
Content
The dataset includes:
3,000+ React components using shadcn UI
Components with Framer Motion animations
Usage examples of Lucide React… See the full description on the dataset page: https://huggingface.co/datasets/valentin-marquez/react-shadcn-codex.gpt-5-codex-250xCodex-Agent-Trace
Codex Contributions to Eyas
This document records the work done by OpenAI Codex on the Eyas pipeline during the Build Small Hackathon.
All Codex commits carry the git co-author trailer Co-Authored-By: Codex <codex@openai.com> and can be verified on GitHub.
Full session-by-session reasoning traces are stored in docs/../codex-traces/ as JSONL files.
What Codex worked on
Multi-camera video grid (All-view)
Built the "All Cameras" view — a synchronized… See the full description on the dataset page: https://huggingface.co/datasets/sehyunlee217/Codex-Agent-Trace.Arcosoph-Codex-Weaver-FC-Reasoning
Arcosoph Codex Weaver Function Calling Reasoning Dataset (V1)
Dataset Description
Welcome to the Arcosoph-Codex-Weaver-FC-Reasoning dataset! This is a comprehensive, multi-source, and meticulously curated dataset designed for instruction-tuning language models to function as intelligent, offline AI agents.
This dataset is provided in a universal, easy-to-parse JSON Lines (.jsonl) format, making it an ideal "source of truth" for creating fine-tuning data for various… See the full description on the dataset page: https://huggingface.co/datasets/arcosoph/Arcosoph-Codex-Weaver-FC-Reasoning.r3al-vit-quantization-codex-trace
R3AL ViT Quantization — Codex Agent Trace
Codex session trace for installing the R3AL CLI and agent skill, exporting
google/vit-base-patch16-224 to ONNX, performing dynamic INT8 post-training
quantization on R3AL, and evaluating model size, Apple-arm64 CPU latency, and
prediction fidelity on a 100-image ImageNet validation sample.
The original Codex JSONL format is preserved for Hugging Face's native Agent
Trace viewer. Credential values, email addresses, unrelated Gmail/Slack… See the full description on the dataset page: https://huggingface.co/datasets/nielsr/r3al-vit-quantization-codex-trace.