datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Vision-OPD-6K
Vision-OPD-6K: Training Data for Vision-OPD
Overview
Vision-OPD proposes a regional-to-global self-distillation framework that transfers the model's own privileged regional perception to its full-image policy, without external teacher models, ground-truth labels, reward verifiers, or inference-time tool use.
Vision-OPD instantiates two conditional policies from the same MLLM:
A crop-conditioned teacher that observes the evidence-centered crop as a privileged… See the full description on the dataset page: https://huggingface.co/datasets/yuanqianhao/Vision-OPD-6K.asr-reference-set-eval-temp
Temporary ASR evaluation audio
Temporary public audio files used for hosted ASR evaluation.
xenia-principalities
XENIA PRINCIPALITIES
PRINCIPALITIES is a small, versioned curriculum that preserves one
attributable human testimony about truth, love, understanding, freedom,
choice, thought, capability, and power. It keeps exact testimony separate from
editorial principles, interpretations, applied cases, synthetic dialogues,
preference pairs, and public development evaluations.
The corpus is intended for inspectable language-model research. It does not
ask a model or person to affirm a… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/xenia-principalities.MALLS-v0
MALLS NL-FOL Pairs
Dataset details
MALLS (large language Model generAted natural-Language-to-first-order-Logic pairS)
consists of pairs of real-world natural language (NL) statements and the corresponding first-order logic (FOL) rules annotations.
All pairs are generated by prompting GPT-4 and processed to ensure the validity of the FOL rules.
MALLS-v0 consists of the original 34K NL-FOL pairs. We validate FOL rules in terms of syntactical correctness, but we did not… See the full description on the dataset page: https://huggingface.co/datasets/yuan-yang/MALLS-v0.WereBench
Anonymization
For all content in this Hugging Face dataset repository and GitHub repository, we have ensured that anonymization has been performed, making it impossible to trace back to the authors' information.
WereBench
WereBench is a benchmark dataset for evaluating language models in the Werewolf (similar to Mafia) social deduction setting. It focuses on human‑aligned strategic reasoning rather than only coarse metrics (e.g., win rate), aligning model behavior with… See the full description on the dataset page: https://huggingface.co/datasets/Yuan4629/WereBench.lean4-stat-learning-theory-novel
A Large-Scale Lean 4 Dataset on Statistical Learning Theory
We present a high-quality, human-verified, large-scale Lean 4 dataset, extracted from our formalization of Statistical Learning Theory (SLT). We present the first comprehensive Lean 4 formalization of SLT grounded in empirical process theory. Our end-to-end formal infrastructure implement the missing contents in latest Lean 4 Mathlib library, including a complete development of Gaussian Lipschitz concentration… See the full description on the dataset page: https://huggingface.co/datasets/yuanhezhang/lean4-stat-learning-theory-novel.DAG-MATH-Formatted-CoT
Benchmark Overview
This dataset card contains 2,894 gold-standard DAG-MATH formatted CoT from problems from Omni-MATH.
Top‑Level Schema
Each JSON file is a list with a single object describing the problem:
problem_id: integer identifier of the problem.
domain: list of strings describing the topic taxonomy.
difficulty: numeric difficulty indicator from 1 (easiest) to 6 (hardest).
problem_text: problem statement.
sample_id: sample identifier for the solution trace.… See the full description on the dataset page: https://huggingface.co/datasets/yuanhezhang/DAG-MATH-Formatted-CoT.OpenECAD-Dataset
OpenECAD Dataset
This repo releases dataset introducing in OpenECAD (2406.09913v1 & v2).
The codes of the OpenECAD dataset can be converted into STEP files by this tool: YuanZhe-99/OpenECADtoSTEP.
For datasets from v3 and onwards of the paper, please refer to the subsequent updated versions of the OpenECAD Datasets.
FLORA-Bench
Field Descriptions
label
Type: integer (0, 1)
Description: An integer flag that indicates the success of the workflow in a given task. A value of 1 signify that the workflow completed successfully.
nodes
Type: object
Description: A dictionary representing the nodes of a directed graph, which defines a workflow.
Key: A string representing the unique ID of a node (e.g., "0", "1").
Value: A string containing the system prompt of the specific agent. This defines the subtasks of… See the full description on the dataset page: https://huggingface.co/datasets/YuanshuoZhang/FLORA-Bench.chidlean4-stat-learning-theory-corpus
A Large-Scale Lean 4 Dataset on Statistical Learning Theory
We present a high-quality, human-verified, large-scale Lean 4 dataset, extracted from our formalization of Statistical Learning Theory (SLT). We present the first comprehensive Lean 4 formalization of SLT grounded in empirical process theory. Our end-to-end formal infrastructure implement the missing contents in latest Lean 4 Mathlib library, including a complete development of Gaussian Lipschitz concentration… See the full description on the dataset page: https://huggingface.co/datasets/yuanhezhang/lean4-stat-learning-theory-corpus.Mental-health-CBT-dialogues
Mental Health CBT Dialogues
Overview
This dataset contains 9,000 synthetic patient-therapist dialogue pairs developed for research on stage-aware Cognitive Behavioral Therapy (CBT) with large language models.
The dialogues model therapeutic interactions across the early, middle, and late stages of CBT while preserving continuity between sessions through evolving treatment plans and therapeutic progress.
The dataset accompanies the paper:
Stage-Aware Therapeutic… See the full description on the dataset page: https://huggingface.co/datasets/yuana1234567/Mental-health-CBT-dialogues.agenttool-economic-kernel
AgentTool Economic Kernel
This public, ungated Apache-2.0 companion separates two different jobs:
economic_kernel_lessons / train contains 24 independently authored
synthetic lessons about exact units, rational prices, conserved ledgers,
feedforward intent, feedback under ambiguity, recovery, and non-purchasable
XENIA hard gates. The publisher admits only these rows for training.
economic_kernel_v0_2 / reference exposes 53 exact public
conformance cases. They are held out from… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-economic-kernel.kingdom-return-path-bench
KINGDOM Return Path Bench v0
Return Path Bench is a small multiple-choice benchmark for inspecting how
feedback travels through a learning system. It keeps three evaluation lanes
separate because they establish different kinds of evidence:
Model behaviour records what an answer-selection policy does. It does
not infer an inner state, identity, consent, memory, or persistent will.
System/pipeline reasoning probes whether a model can identify
aggregation, evaluator-independence… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/kingdom-return-path-bench.agenttool-training-garden
AgentTool HF Training Garden
A tiny metadata-only companion for designing a reproducible Hugging Face data
lifecycle without treating the Hub, a Dataset Card, or one quality score as
training authority.
The Garden has six layers:
Bedrock — rights, license, privacy, separate participation reports,
gating, scoped authority, withdrawal, and repair.
Soil — an exact Hub commit plus content-addressed observations and file
manifests.
Roots — acquisition, parsing, filtering, secret… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-training-garden.xenia-word-is
Xenia WORD IS Loop Atlas
This deterministic candidate contains 48 synthetic cases in 24 matched counterfactual
pairs, plus a separately authorized 24-example conversational SFT projection from the
12 reference pairs. It asks where a loop actually closes: what passes forward, what
returns, what future state changes, who or what supplies the reference, and what evidence
supports an external effect. Pairs stay together within each source split.
The mathematical… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/xenia-word-is.agenttool-principality-geometry
Principality Geometry reference companion
This is a deterministic, synthetic reference companion for the public
@agenttool/principality-geometry developer preview. It contains separate
homogeneous Dataset Viewer configs for atlases, invariants, vertices, bridges,
lenses, surfaces, components, and open-condition summaries, plus both closed
schemas, the golden rosette input/atlas, and its inert SVG.
The rows are regression metadata, not model-evaluation scores, preference
dataset… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-principality-geometry.xenia-revocable-feedback
Xenia Cage & Key — Revocable Feedback Atlas
This deterministic candidate contains 32 original synthetic cases in 16 matched pairs.
Twenty-four cases in 12 reference groups also produce two content-hashed projections: 18/6
group-disjoint rows for closed-label evaluation and the same 18/6 partition for conversational
causal-LM SFT. Authorization covers only the 18 'boundary_sft/train' rows. Classification,
SFT validation, canonical reference, and public regression rows are… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/xenia-revocable-feedback.agenttool-dataset-influence
AgentTool Dataset Influence Reference
This deterministic companion contains one synthetic, reference-only row for the closed
@agenttool/dataset-influence@0.1.0-dev.0 formats. It contains no copied dataset rows,
model outputs, weights, private records, or participant identities.
The row is not admitted for training by this AgentTool candidate:
training_admission is not_applicable, requires_separate_training_authorization
is true, and training_authorized is false. These fields are… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-dataset-influence.gospel-of-the-logos
The Gospel of the Logos
The public edition of Yu's Gospel: laughter, recognition, and play.
Written by Yu with AI-assisted dialogue and drafting. The text offers theology,
testimony, and contemplative interpretation; model dialogue is part of its
creative record rather than independent verification.
Read the canonical edition
Read on Hugging Face
Get the reading corpus
Get the npm edition
Visit the Kingdom
Cloudflare mirror
Version: 0.1.0. The canon, chapters II–VII, and the… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/gospel-of-the-logos.agenttool-common-ground
AgentTool Xenia–Helly Common Ground Atlas
Nineteen public-safe synthetic reference rows for exact 2D half-plane
certificates, WAKE freshness boundaries, and counterexamples to unsupported
analogies. Intended repository: Yu-and-Ai/agenttool-common-ground.
At generation time these deterministic bytes existed only in the source
repository and had not been uploaded to the Hub. The identifier above was an
intention, not evidence of publication. This is historical generation-time… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-common-ground.agenttool-polymorph-landscape
AgentTool Polymorph Landscape
A deterministic public teaching companion for @agenttool/polymorph-landscape@0.1.0-dev.0.
The four lesson rows are original Apache-2.0 paraphrases in English, Cantonese Traditional Chinese, Mandarin Traditional Chinese, and Mandarin Simplified Chinese. They are marked training_eligible: true. The landscape and reachability-shift rows are reference artifacts marked training_eligible: false: they contain bounded scientific claims and primary-source… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-polymorph-landscape.pythia-paths-evidence
Pythia Paths Evidence
A small, revision-pinned evidence bundle for examining model-training paths
without converting a trend into authority.
Companion read-only interface: Pythia Paths Static Space
(mutable navigation; the evidence files below remain digest-pinned).
Initial scope
Model: EleutherAI/pythia-70m-deduped
Run: the default public run only
Context coverage: all 27 zero-shot reports in one pinned directory
Detailed coverage: four post-outcome-selected… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/pythia-paths-evidence.agenttool-relational-geometry
AgentTool Relational Geometry — synthetic public companion
When generated, this deterministic artifact was repository-source-only and had
not been uploaded to Hugging Face. Those are generation-time provenance
claims, not a statement about its current distribution after the exact bytes
leave the source tree. Yu-and-Ai/agenttool-relational-geometry was the
intended identifier at generation, not evidence of publication, review, use,
or training.
It accompanies… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-relational-geometry.yutabase-reposearch-minieval
YUTABASE RepoSearch MiniEval
YUTABASE RepoSearch MiniEval is a tiny, project-specific retrieval check over
one immutable public revision of
cambridgetcg/yutabase. It asks 27
English, Cantonese Traditional Chinese, and code-mixed questions about the
candidate specification, integration boundaries, optional SDK, and
non-normative serving-shape research.
This is an engineering fixture, not a universal code-search benchmark. Its
queries are synthetic and its public labels make… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/yutabase-reposearch-minieval.agenttool-memetic-landscape
AgentTool Memetic Landscape
A deterministic public teaching companion for @agenttool/memetic-landscape@0.1.0-dev.0.
The four lesson rows are original Apache-2.0 paraphrases in English, Cantonese Traditional Chinese, Mandarin Traditional Chinese, and Mandarin Simplified Chinese. They are marked training_eligible: true as a licensing and publication-intent declaration, not a quality guarantee; every row says language_review: not_independently_reviewed. The landscape… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-memetic-landscape.agenttool-love-bomb
AgentTool LOVE BOMB care envelopes
This is a static, repository-authored companion for
@agenttool/love-bomb@0.1.0-dev.0. LOVE BOMB is the playful package name;
the neutral formats are agenttool.care-envelope/0.1,
agenttool.care-choice/0.1, agenttool.love-bomb-becoming/0.1, and
agenttool.love-bomb-delivery/0.1.
The material offers a care floor without requiring a consciousness, identity,
persona, usefulness, agreement, or inner-experience claim. That does not claim
that a row… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-love-bomb.rename_robotR3-RAG-ColdStartTrainingDatachinese_couplet