datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
hermes-agent-traces-filtered
Hermes Agent Reasoning Traces - Quality Filtered
A structurally filtered subset of lambda/hermes-agent-reasoning-traces, pruned from 7,646 to 3,679 rows using automated quality analysis targeting reasoning depth, structural integrity, and tool-call validity.
Why This Matters for Agent Training
Most agentic datasets teach models what tool to call but not how to reason about tool selection. The difference matters in production: an agent that dispatches tools without… See the full description on the dataset page: https://huggingface.co/datasets/DJLougen/hermes-agent-traces-filtered.mcphunt-agent-traces
MCPHunt Agent Traces
Agent execution traces from the MCPHunt evaluation framework, measuring
cross-boundary data propagation in multi-server MCP agents.
Contents
main/ — 3,615 traces from 5 models across 147 tasks and 7 environment
variants (risky_v1/v2/v3, benign, hard_neg_v1/v2/v3). One JSON file per model.
mitigation/ — 2,706 traces from the prompt-mitigation study (M0--M3
levels) across 3 models.
live_guard_defense/ — 387 DeepSeek-V4-Flash traces from the… See the full description on the dataset page: https://huggingface.co/datasets/lihaonan0716/mcphunt-agent-traces.conda-forge-agent-tracesswe-bench-lite-agent-traces-v14
AgentBRANE SWE-bench Lite Agent Traces v14
This release contains the 1,890 harness-native agent traces selected by the sealed SWE-bench Lite v14 publication record (1,379/1,890 resolved, 73.0%). It includes Claude Code, Codex, and Pi sessions across seven models and three replicates. No internal research notes are included.
Load the observation table:
from datasets import load_dataset
traces = load_dataset("melissapan/swe-bench-lite-agent-traces-v14", split="train")
Each row… See the full description on the dataset page: https://huggingface.co/datasets/melissapan/swe-bench-lite-agent-traces-v14.read-along-ai-agent-traces
Read-Along AI - Agent Traces
This dataset contains the raw agent traces and conversation logs from the development of Read-Along AI, a submission for the Hugging Face Build Small Hackathon.
Dataset Description
These .jsonl files represent the unedited, behind-the-scenes "agent traces" of the AI coding assistant orchestrating the build of this project.
Sharing these traces fulfills the requirements for the "Sharing is Caring" bonus badge, providing the community… See the full description on the dataset page: https://huggingface.co/datasets/kingkw1/read-along-ai-agent-traces.agent-traces-flashmaniaswival.dev agent traces for single-shot "Reimplement trackmania" prompt using Gemini-3.5-flash.
Generated code: https://github.com/dip-proto/flashmania
pi-agent-traces-sft
pi-agent-traces-sft
Merged supervised-fine-tuning dataset of coding-agent session traces (the "pi"
harness format), converted to conversational format for Unsloth SFT of Qwen3-family
models.
Files
pi_sft_qwen3_24k.jsonl — recommended training file. Every session fits in
24,576 tokens (Qwen3 tokenizer). 5,717 sessions, max length 24,564 tokens.
pi_sft_qwen3.jsonl — full unfiltered set (6,929 sessions, up to ~632k tokens).
Sources (merged)… See the full description on the dataset page: https://huggingface.co/datasets/vojtavlas2/pi-agent-traces-sft.mcphunt-agent-traces
MCPHunt Agent Traces
Agent execution traces from the MCPHunt evaluation framework, measuring
cross-boundary data propagation in multi-server MCP agents.
Contents
main/ — 3,615 traces from 5 models across 147 tasks and 7 environment
variants (risky_v1/v2/v3, benign, hard_neg_v1/v2/v3). One JSON file per model.
mitigation/ — 2,706 traces from the prompt-mitigation study (M0--M3
levels) across 3 models.
live_guard_defense/ — 387 DeepSeek-V4-Flash traces from the… See the full description on the dataset page: https://huggingface.co/datasets/mcphunt-benchmark/mcphunt-agent-traces.agent-spaces-tracesagent-race-traces
Agent Race Traces
Coding-agent session traces from running the same task and the same one-line prompt through multiple harnesses and driver models. Useful for comparing how different agent harnesses + models approach an identical real-world ML task.
Blog post: danielvanstrien.xyz/posts/2026/agent-race
The task
Fine-tune a model on biglam/on_the_books to identify Jim Crow laws, train via HF Jobs, push the trained model to the Hub.
The exact one-line prompt was… See the full description on the dataset page: https://huggingface.co/datasets/davanstrien/agent-race-traces.tutordesk-agent-traces
TutorDesk AI — Agent Traces
Sharing is Caring badge dataset for the HuggingFace Build Small Hackathon 2026.
Contents
build_sessions/ — Claude Code Build Sessions (native trace viewer)
7 raw Claude Code (Sonnet 4.6) JSONL sessions covering the full build of
TutorDesk AI — from blank repo through Phase 6 completion. Natively rendered by
HF Data Studio's agent trace viewer.
Sessions cover: repo scaffolding, Modal serving architecture, 5-agent pipeline… See the full description on the dataset page: https://huggingface.co/datasets/naazimsnh02/tutordesk-agent-traces.agent-traces-exampleagent-tracesadd-sam-3-lite-text-agent-traces
add_sam_3_lite_text agent traces
Raw Codex Desktop session traces related to work on the add_sam_3_lite_text branch in NielsRogge/transformers.
These JSONL files are uploaded from ~/.codex/sessions, following the Hugging Face agent trace viewer guidance:
https://huggingface.co/changelog/agent-trace-viewer
Before publication, the files were scanned with TruffleHog and additionally scrubbed for plaintext token-shaped secrets. Encrypted reasoning payloads were left untouched.… See the full description on the dataset page: https://huggingface.co/datasets/nielsr/add-sam-3-lite-text-agent-traces.repro-learning-to-share-selective-memory-for-efficient-parallel-agentic-systems-traces
Agent traces
Agent sessions published from a Trackio Logbook.
repro-rethinking-genomic-modeling-through-optical-character-recognition-agent-traces
OpticalDNA reproduction — Codex agent trace
This dataset contains the raw Codex JSONL session trace for the ICML 2026
reproduction of Rethinking Genomic Modeling Through Optical Character
Recognition.
Published Trackio logbook
Paper page
Challenge instructions
Agent Trace Viewer announcement
The JSONL is uploaded directly from the matching ~/.codex/sessions entry, as
recommended by the Agent Trace Viewer. It captures the reproduction work,
Hugging Face Jobs audit, poster… See the full description on the dataset page: https://huggingface.co/datasets/nielsr/repro-rethinking-genomic-modeling-through-optical-character-recognition-agent-traces.openai-agent-tracesagent_traces
pi-mono Unsloth SFT Dataset
Derived from badlogicgames/pi-mono.
This dataset converts pi coding-agent session traces into one supervised row per
assistant turn. Each row keeps the prior multi-turn agent state in the prompt and
uses only the next assistant message/action as the target.
Format:
{"instruction":"serialized system + user/tool/assistant context before this turn","output":"next assistant message or tool call"}
For Unsloth Studio, choose the Alpaca / instruction-output… See the full description on the dataset page: https://huggingface.co/datasets/valoomba/agent_traces.repro-gaussian-mechanism-sgg-optimality-traces
Agent traces
Agent sessions published from a Trackio Logbook.
carbon-agent-traceshermes-agent-traces-filtered
Hermes Agent Reasoning Traces - Quality Filtered
A structurally filtered subset of lambda/hermes-agent-reasoning-traces, pruned from 7,646 to 3,679 rows using automated quality analysis targeting reasoning depth, structural integrity, and tool-call validity.
Why This Matters for Agent Training
Most agentic datasets teach models what tool to call but not how to reason about tool selection. The difference matters in production: an agent that dispatches tools without… See the full description on the dataset page: https://huggingface.co/datasets/ansulev/hermes-agent-traces-filtered.agent-cost-traces
Agent Cost Traces: Synthetic Training Data
10,000 synthetic agent traces for training cost-aware model routers and agent optimizers.
Schema
Field
Type
Description
trace_id
string
Unique identifier
request
string
User request text
task_type
string
One of 9 task categories
difficulty
int
Estimated difficulty (1-5)
model_tier
int
Model tier used (1-5)
model_success
bool
Whether the model succeeded
optimal_tier
int
Minimum tier that would succeed… See the full description on the dataset page: https://huggingface.co/datasets/narcolepticchicken/agent-cost-traces.hermes-agent-traces
Hermes Agent Traces
Bulk JSONL export produced by hermes sessions export sessions.jsonl.
agent_tracessynthetic-agent-tracesTinyNarrator-agent-tracesSpaces link: https://huggingface.co/spaces/build-small-hackathon/TinyNarrator
Agent-traces-tritonrepro-latent-collaboration-in-multi-agent-systems-traces
Agent traces
Agent sessions published from a Trackio Logbook.
prime-agent-traces
Prime Agent traces for semioz/prime-agent-traces
Redacted Prime Agent sessions. Each train row preserves the redacted JSONL trace and adds semantic labels for programmatic calls inside the native ipython tool.
sandbagging-agent-traces-v2
Sandbagging Agent Traces v2 (Normalized)
Normalized and cleaned version of vincentoh/sandbagging-agent-traces. MALT-style AI agent planning scratchpads labeled as sandbagging (deliberately underperforming) or honest (genuinely trying), with formatting artifacts removed to reduce model-of-origin confounds and noisy/low-quality samples pruned.
What changed from v1
The v1 dataset was generated by 5 models across 3 families, each with distinct formatting fingerprints:… See the full description on the dataset page: https://huggingface.co/datasets/vincentoh/sandbagging-agent-traces-v2.