datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
karate-wiener-animationsfireboy-vla-rollout-artifacts
Fire Boy VLA Rollout And Evidence Artifacts
Creator And Submission Links
Item
Link
Creator
Sanjay Prasad H S (sanjuhs)
GitHub repo
sanjuhs/build-small-hackathon-v1
Canonical HF Space repo
build-small-hackathon/toy-room-v3
Canonical live Space
https://build-small-hackathon-toy-room-v3.hf.space/toy-v3
Personal HF Space mirror
sanjuhs/toy-room-v3
Personal live Space mirror
https://sanjuhs-toy-room-v3.hf.space/toy-v3
YouTube demo… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/fireboy-vla-rollout-artifacts.agenda-parser-tool-traces
Agenda Parser — tool-calling reasoning traces
ReAct tool-calling traces for the Agenda Parser
agents: each row is one agent step — a {system, user, assistant} chat example
where the assistant emits a single JSON action {"thought", "tool", "args"}.
Two agents are covered (tagged by meta.domain):
agenda — the uploaded-packet research agent, over real public-meeting agenda
packets (tools: list/read items, semantic + exact search, summarize, report).
Each agenda row's meta.unit_id… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/agenda-parser-tool-traces.jawbreaker-scam-defense-data
Jawbreaker Scam Defense Data
Synthetic and sanitized training/eval data for Jawbreaker, a local-first scam defense app for someone you love.
Jawbreaker turns a suspicious text, email, or DM into a plain-English safety card: the risk, the warning signs, and the safest next step before someone replies, clicks, or pays.
Contents
eval/: scam-defense evaluation sets from smoke checks through hard calibration suites.
eval/reports/: guarded evaluation reports for the… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/jawbreaker-scam-defense-data.build-small-shadowing-mini-audio
Shadowing Practice Mini Reference Audio
Prebuilt starter reference audio for the Language Practice Lab hackathon Space.
Files are short WAV examples used to make the demo responsive before live synthesis is needed.
kicky-ai-spf
FUT-HEROS SPF — COCO instance-segmentation dataset (ball / player / goal)
Auto-labelled football frames (COCO format) for ball/player/goal instance segmentation.
Labels generated by SAM3 + NVIDIA LocateAnything-3B (zero manual annotation), from a
single fixed-camera amateur session. Clip-level, goal-stratified train/valid/test split.
train/, valid/, test/ each have _annotations.coco.json + frames.
Classes: 1 ball, 2 player, 3 goal. Part of the FUT-HEROS project.
ux-crime-scene-traces
🔎 UX Crime Scene — Investigation Traces
Real agent traces from UX Crime Scene,
a film-noir detective that investigates UI screenshots as crime scenes — built for the
Build Small Hackathon (Gradio × Hugging Face).
Each row is one real investigation: the input screenshot the user dropped, and the
raw structured verdict Qwen2.5-VL-7B returned — the crimes it found, the bounding
box of each guilty element, the testimony, the severity, and the final grade.
The set spans different… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/ux-crime-scene-traces.CVE_Vulnerailities_Detailedagent-trace-privacy-scrubber-codex-tracesdota2tuned-data
DOTA2Tuned Data
This dataset supports the DOTA2Tuned Hugging Face Build Small Hackathon app. It contains compact derived artifacts for Dota 2 draft recommendations, hero meta lookup, build timing summaries, match prediction, retrieval, and supervised fine-tuning examples.
Contents
sft_examples.jsonl: instruction examples generated from normalized Dota 2 recommendations, patch/stat cards, and app behaviors.
Compact Parquet artifacts used by the Space:
dim_hero… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/dota2tuned-data.lolaby-traces
Lolaby — generation traces
Pipeline traces from Lolaby, an AI-powered lullaby generator built for the Build Small Hackathon 2026 (Backyard AI track).
Each trace is a complete witness of one end-to-end generation: every input the user gave, every model that ran, every prompt and raw output, every timing measurement, and the final audio. Published under CC0 so anyone can study, replay, or remix the pipeline.
What's in a trace
Each subfolder is one generation. Files:… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/lolaby-traces.kirana-detective-build-traces
Kirana Detective — Claude Code Build Sessions
Raw Claude Code (claude-sonnet-4-6) session traces recorded while building
Kirana Detective AI for the HuggingFace Build Small Hackathon 2026.
Each .jsonl file is one coding session. Together they cover the entire
build — from first commit to final submission.
What's Inside
Sessions
Agent
Coverage
11 JSONL files
Claude Code (Sonnet 4.6)
Full project build
Sessions include
Designing the… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/kirana-detective-build-traces.discoverroute-citieshackathon-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.pit-wall-chaos-tracesCodex agent traces for Pit Wall Chaos, a Build Small Hackathon project.
Space link: https://huggingface.co/spaces/build-small-hackathon/pit-wall-chaos
job-search-assistant-agent-tracemafia-dataset
Mafia Dataset
Design package for a unified Mafia-agent fine-tuning and evaluation dataset.
Target game:
7 players
2 Mafia
1 Detective
1 Doctor
3 Villagers
day/night cycles
public discussion
role-aware Time-to-Talk communication
classic win conditions: Mafia win at parity or majority; town wins when all Mafia are eliminated
This folder is intentionally split into contracts, source registry, examples, and audits. Raw upstream data stays in the original repos and source folders.… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/mafia-dataset.figment-eval-traces
Figment Eval Traces
Synthetic and de-identified evaluation traces for Figment, a prototype protocol-navigation aid for trained rural-clinic and disaster-response field responders.
These records are intended for model and harness debugging. They are not clinical data, medical advice, diagnosis, treatment instructions, or a substitute for local protocol, clinician judgment, supervisor review, or trained responder judgment.
Dataset Summary
The dataset captures… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/figment-eval-traces.TinyNarrator-agent-tracesSpaces link: https://huggingface.co/spaces/build-small-hackathon/TinyNarrator
NeuroBait-Codex-Traces
Codex Session Traces
This folder contains Codex rollout JSONL traces related to the
NeuroBait Build Small Model project.
Included traces:
rollout-2026-06-08T17-03-23-019ea6af-db29-7801-ac01-46dfc88f90b0.jsonl
rollout-2026-06-09T07-10-21-019ea9b7-4610-7223-906e-2d0dba8bae7f.jsonl
rollout-2026-06-09T16-00-29-019eab9c-a18a-7de1-8967-ea63db425a4f.jsonl
The traces were selected because their session metadata contains the project
working directory:… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/NeuroBait-Codex-Traces.build-small-agent-trace
Codex agent trace
This is submitted as part of the HF Build Small hack.
Codex was fully used for building the project.
It is necessary to share the trace to be eligible for the OpenAI price.
Trace was screened and sensitive keys redacted before sharing.
The full running app can be seen at :
https://huggingface.co/spaces/build-small-hackathon/storyboard-tui
small-signalsdiscoverroute-tracesMatchWise-agent-tracemind-of-tashi-selfplay
The Mind of Tashi — self-play traces
Self-play data for SFT of a small reasoning model that plays The Mind of
Tashi — a simultaneous-commit ritual fighting game where the opponent's
<think> block is the game (surfaced to the player as the "mind-scroll").
Two LLMs duel each other under the game's blind-commit contract (each side
sees only the match history, never the opponent's pending move); we keep the
opponent side's full <think> + {move, taunt} as the SFT target.
Part of the… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/mind-of-tashi-selfplay.best-man-speech-codex-traces
Best Man Speech Codex Traces
This dataset contains a sanitized Codex agent trace from development work on Best Man Speech
Coach, a Gradio app built for the Hugging Face Build Small Hackathon.
The trace captures a real software engineering session involving repository inspection,
implementation planning, code edits, validation checks, GitHub pull request context, and follow-up
project hygiene around sharing agent traces.
Contents… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/best-man-speech-codex-traces.in-your-own-worlds-tracesCodex agent traces for In your own wor(l)ds, a Build Small Hackathon project.
Space link: https://huggingface.co/spaces/build-small-hackathon/in-your-own-worlds
ai-prophecy-court-presence
AI Prophecy Court Presence
Exploration-ready normalized records for AI Prophecy Court, a playful
hackathon project examining the public statements and social presence of major
AI leaders.
Dataset configurations
linkedin: original authored LinkedIn posts from four verified profiles
x: posts, replies, quotes, and visible reposts from six verified profiles
Each row retains its source URL, publication time, content type, engagement
metadata, collection ID, run ID… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/ai-prophecy-court-presence.kirana-invoice-train-data
Kirana Invoice Training Data — Indian FMCG
Training dataset for the Kirana Detective project — an AI pipeline that audits distributor invoices for Indian kirana (grocery) stores. The repository contains two distinct sub-datasets used to fine-tune two separate models.
Dataset Summary
Sub-dataset
Purpose
Size
Format
synthetic_invoices/
OCR fine-tuning (MiniCPM-V)
500 images + annotations
PNG + JSONL
fmcg_catalog.json
Product name normalization… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/kirana-invoice-train-data.moku-blog-assets