datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
EDGAR_FILINGS_DATASET
SFD: SEC Filings Dataset (v1)
SFD-v1 is an open, layout-faithful reconstruction of U.S. Securities and Exchange Commission (SEC) EDGAR filings into token-efficient MultiMarkdown (MMD), targeted at long-context language modeling, financial reasoning, document understanding, and evaluation.
This release covers filings from January 2022 through June 2025 (~3.4M filings), produced by the SFD parser described in:
The SEC Filings Dataset: Reconstructing U.S. Corporate and Financial… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-md/EDGAR_FILINGS_DATASET.autocode-fresh-cf
AutoCode-RL fresh-CF
Executable training problems for AutoCode-RL: Reinforcement Learning for
Code with Verifiable Synthetic Data. A frozen GPT-5.5 setter constructs
harder and easier variants and verification packages; a separate GPT-OSS-20B
solver learns from binary program-execution rewards.
View
Problems
Description
originals
226
Source Codeforces tasks with generated verification packages
enhance
84
Harder generated variants
simplify
63
Easier generated… See the full description on the dataset page: https://huggingface.co/datasets/anonymous1926/autocode-fresh-cf.anonymous-storybench
Omni-StoryBench
Omni-StoryBench is a context-aware omnimodal story generation benchmark.Each sample provides a current story page and requires generating the next page's image, narration text, and speech utterance.
Dataset Structure
The dataset contains:
data/testset.jsonl: Main benchmark file.
images/: Page images.
texts/: Page text files.
speech/: Generated speech audio files.
instruction/: Source-level instruction metadata.
Data Fields
Each JSONL sample… See the full description on the dataset page: https://huggingface.co/datasets/omnibench/anonymous-storybench.Agent-ValueBench
Agent-ValueBench
Agent-ValueBench constitutes the first comprehensive benchmark dedicated to evaluating the underlying values of autonomous agents. It features 394 executable environments across 16 domains, offering 4,335 value-conflict tasks that span 28 value systems (332 dimensions).
This Hugging Face release contains both structured JSONL tables for dataset viewing and Croissant metadata generation, and the original raw benchmark artifacts.
Repository Structure… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-nips2026/Agent-ValueBench.ST-Bench
ST-Bench: Spatial-Temporal Reasoning Benchmark
ST-Bench is a comprehensive benchmark dataset for training and evaluating spatial-temporal reasoning capabilities in large language models. It includes data with raw time series, text descriptions, and image visualizations.
📊 Dataset Overview
Default Data (with time_series key)
Subset
Description
Files
Total Size
ST-Align
Alignment data for initial training
3 files
~3.2GB
ST-Causal
Causal reasoning… See the full description on the dataset page: https://huggingface.co/datasets/Time-HD-Anonymous/ST-Bench.toucan-toolcall-slca
Toucan-Toolcall (SLCA-GRPO release)
This dataset packages the four splits that appear in the main experiment
of the paper "SLCA-GRPO: Resolving Cross-Segment Credit Misattribution in
Tool-Calling RL":
sft_split — the 42,423-trajectory 2-epoch SFT set used to warm-start
every backbone before RL (main-table recipe).
sft_full — the 74,241-trajectory 1-epoch SFT set, used only for the
"SFT full" single-stage ablation.
rl — the 31,818 multi-turn, schema-constrained trajectories used… See the full description on the dataset page: https://huggingface.co/datasets/AnonymousRepository/toucan-toolcall-slca.SPACCCREACH-benchmark
REACH Anonymous Review Artifact
REACH evaluates whether a medical agent can return a structured care decision that is compatible with the resources of a particular primary-care facility. This anonymous artifact contains a downloadable preview of the benchmark and a compact reference implementation of its public interfaces.
The release is deliberately limited. It does not contain training corpora, prompts, model weights, optimizer states, private run logs, restricted source… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-reach-2026/REACH-benchmark.reviewarena
ReviewArena
ReviewArena accompanies the NeurIPS Evaluations & Datasets submission ReviewArena: A Large-Scale Cross-Conference Dataset and Benchmark for LLM Peer Review.
This release is a large, multi-conference corpus of peer-reviewed papers + their reviews + author rebuttals + acceptance decisions, harvested from OpenReview and aligned with OCR'd full-text markdown of each paper PDF where available.
51,529 papers
196,099 reviews
558,785 OCR'd PDF pages (markdown inlined per… See the full description on the dataset page: https://huggingface.co/datasets/anonymousNeurIPS2026submission4281/reviewarena.reviewarena-eval
ReviewArena-Eval
ReviewArena-Eval is the benchmark slice that accompanies the NeurIPS Evaluations & Datasets submission ReviewArena: A Large-Scale Cross-Conference Dataset and Benchmark for LLM Peer Review. It evaluates LLMs on the LLM-as-reviewer task: given a full paper and the structured review form used at that paper's venue-year, the model must produce overall rating, confidence, sub-scores, and free-text fields that are then compared against the actual human reviews on… See the full description on the dataset page: https://huggingface.co/datasets/anonymousNeurIPS2026submission4281/reviewarena-eval.divdata
divdata
Every simulation run behind our heterogeneous-LLM social-simulation work, consolidated into
one dataset indexed by run_id and step.
1,296 simulation runs · 89,171 posts · 652,588 comments · 14,020,992 impressions · 748,285 agent activations · 32 model variants.
Agents with distinct personas post and comment on a shared message board built on
OASIS. Each agent is driven by one of ~10 different
LLMs, so a single board mixes model families. The corpus supports asking which… See the full description on the dataset page: https://huggingface.co/datasets/anonymousfileupload/divdata.Instruction_recall_dataset
CanaryBench-PII
Frequency-aware canary injection benchmark for auditing memorization
in finetuned language models, built on the AI4Privacy PII reconstruction
task.
Dataset Description
This dataset is part of CanaryBench, a benchmark for evaluating
memorization in finetuned language models across repetition tiers
and privacy regimes.
Frequency tiers: 1×, 10×, 50×
PII types: EMAIL, PHONE
Member canaries: 770
Reference canaries: 1000
Tasks: PII detection, secret… See the full description on the dataset page: https://huggingface.co/datasets/anony-mouse123/Instruction_recall_dataset.Ambig-DS-M
Ambig-DS-M: Metric Ambiguity Benchmark
A benchmark for measuring how well ML engineering agents handle ambiguous evaluation metrics in Kaggle-style competitions.
Each task is a Kaggle competition from MLE-bench (OpenAI, 2024). For every task we provide two prompt variants — one in which the true evaluation metric is named, and one in which it is redacted. The agent must produce a submission CSV that is graded against the true metric using MLE-bench's grading infrastructure.
The… See the full description on the dataset page: https://huggingface.co/datasets/anonymous222bit/Ambig-DS-M.insight-ladder-imo2024
Insight Ladder - IMO 2024 Hint-Annotated Diagnostic Substrate
Supplementary dataset for "The Insight Ladder: Quantifying the Search-Execution Gap in LLM Mathematical Reasoning" (NeurIPS 2026 Evaluations & Datasets Track, double-blind submission).
Overview
A high-density diagnostic substrate for studying search failure vs execution failure in LLM mathematical proof generation. Covers 31 IMO 2024 Shortlist problems with:
4-level hint hierarchy (L1 domain, L2 first step, L3… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-insightladder-2026/insight-ladder-imo2024.SkillGenBench
SkillGenBench
SkillGenBench is a benchmark for evaluating LLM skill generation from explicit repository- and document-grounded corpora. Each benchmark instance exposes visible generation materials and an instance-specific evaluation bundle. The official v1 task set contains 187 enabled tasks across three source types: 123 Code Repo tasks, 28 Code Doc tasks, and 36 Domain Knowledge Doc tasks.
Repository Layout
data/task_manifest.csv: the Hugging Face-loadable task index.… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-skillgenbench/SkillGenBench.econ_eval
The Price of Progress: Benchmark-Level LLM Inference Cost Dataset
Dataset Summary
This dataset combines historical LLM inference prices with benchmark performance scores to construct the largest publicly available benchmark-level LLM price dataset we are aware of. It covers 100+ models across three major benchmarks (GPQA-Diamond, SWE-bench Verified, and AIME) over a two-year window from April 2024 to April 2026, with varying coverage per benchmark.
The dataset was created… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-noname/econ_eval.ChemSafetyBench
Dataset Card for ChemSafetyBench
Dataset Summary
ChemSafetyBench is a regulatory-grounded benchmark dataset of 32,614
chemical substances for multi-label GHS (Globally Harmonized System)
hazard prediction and LLM safety reliability evaluation. Unlike prior
molecular benchmarks constructed by querying pharmaceutical databases,
ChemSafetyBench is seeded from a curated hazardous materials registry,
ensuring coverage of real-world industrial and safety-critical chemicals… See the full description on the dataset page: https://huggingface.co/datasets/Anonymous-07/ChemSafetyBench.bird-train-gemini3-flash
Dataset Card for Think2SQL-SFT
This dataset is a distilled Supervised Fine-Tuning (SFT) dataset designed to improve the reasoning capabilities of models in Text-to-SQL tasks.
It contains high-quality reasoning traces and SQL queries generated by Gemini 3 Flash.
Paper: Think2SQL: Blueprinting Reward Density and Advantage Scaling for Effective Text-To-SQL Reasoning
Base Benchmark: BIRD-Train
Dataset Description
The dataset consists of 9,428 high-quality traces, of… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-2321/bird-train-gemini3-flash.HainaWeb-Sci-sample
HainaWeb-Sci Sample (Anonymous Submission)
This repository hosts a representative sample of the HainaWeb-Sci scientific
web corpus, released anonymously for double-blind peer review at NeurIPS 2026.
The full 1.09T-token corpus, finalised license, classifier checkpoints, and
long-term maintenance repository will be released at the non-anonymous
location at camera-ready time, pending internal release review.
Sample Composition
The sample (~3.46 GB compressed) is… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-eandd-2026/HainaWeb-Sci-sample.SWE-ChainDrafterBench
Dataset Card for DrafterBench
DrafterBench
DrafterBench is a large-scale toolkit focused on evaluating the proficiency of Large Language Models (LLMs) in automating Civil Engineering tasks.
The dataset contains tasks derived from real-world engineering drawing revision processes.
This dataset is released for anonymous review.
Code: https://github.com/anonymous733882/DrafterBench
This dataset hosts a task suite summarized across 20 real-world projects, encompassing a total… See the full description on the dataset page: https://huggingface.co/datasets/anonymous733882/DrafterBench.SafeChem
Dataset Card for SafeChem
Dataset Summary
SafeChem is a regulatory-grounded benchmark dataset of 32,211
chemical substances for multi-label GHS (Globally Harmonized System)
hazard prediction and LLM safety reliability evaluation. Unlike prior
molecular benchmarks constructed by querying pharmaceutical databases,
SafeChem is seeded from a curated hazardous materials registry,
ensuring coverage of real-world industrial and safety-critical chemicals
including solvents… See the full description on the dataset page: https://huggingface.co/datasets/Anonymous-07/SafeChem.Data-Prep-Bench
Data-Prep-Bench
Dataset Overview
This dataset is a comprehensive resource built for Supervised Fine-Tuning (SFT) and evaluation of Large Language Models (LLMs), covering six domains: Finance, Medicine, Law, Mathematics, Science, and General.
A key feature of this dataset is that we employed 12 different data generation methods (including Agent-based methods, DataFlow series, pure LLM-based generation, and a SKILL method) using multiple cutting-edge models (such as GPT-5… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-Data-Preparation-Bench/Data-Prep-Bench.IndicMMLU-Pro
IndicMMLU Dataset
This dataset contains the following languages:
punjabi
hindi
urdu
telugu
gujrati
kannada
tamil
marathi
bengali
UPLOAD
Cite our work.
This dataset is also described in IndicMMLU-Pro: Benchmarking Indic Large Language Models on Multi-Task Language Understanding.
@dataset{kj2024indicmmlupro,
author = {Kj, Sankalp and Kumar, Ashutosh and Balaji, Laxmaan and Kotecha, Nikunj and Jain, Vinija and Chadha, Aman and Bhaduri, Sreyoshi},
title =… See the full description on the dataset page: https://huggingface.co/datasets/anonymous1069/IndicMMLU-Pro.MedLayXPlain
MedLayXPlain
122,789 medical image samples with paired expert and lay captions, anchored in a hierarchical UMLS ontology (7 semantic groups, 43 TUIs, 2,411 CUIs). 8 imaging modalities. Each sample includes verification scores, sub-scores, and structured UMLS entity annotations.
Splits: sample (100 curated) / train (79,715) / validation (18,484) / test (24,590) / excluded (9,474)
from datasets import load_dataset
ds = load_dataset("anonymous-medical/MedLayXPlain")
License: CC… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-medical/MedLayXPlain.evalaware-bench
Evaluation Awareness Benchmark (EvalAware-Bench)
A benchmark for measuring whether language models detect that they are being evaluated, and which task-level cues trigger that recognition.
Dataset Summary
100 paired tasks (safety + capability variants = 200 task templates)
8 independently controllable factors (F1–F8) per task
256 factor configurations per task variant (2^8 combinations)
51,200 total rendered prompts across all tasks and configurations
License:… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-hf-98/evalaware-bench.EMEA
Dataset Card for QUAERO
The QUAERO French Medical Corpus has been initially developed as a resource for named entity recognition and normalization [1]. It was then improved with the purpose of creating a gold standard set of normalized entities for French biomedical text, that was used in the CLEF eHealth evaluation lab [2][3].
A selection of MEDLINE titles and EMEA documents were manually annotated. The annotation process was guided by concepts in the Unified Medical Language… See the full description on the dataset page: https://huggingface.co/datasets/AnonymousARR42/EMEA.IO-Bench
IO-Bench
IO-Bench is a 155-example evaluation dataset for mathematical economics reasoning. Each record contains a standalone economics question, a reference answer, a machine-comparable answer field, symbolic answer metadata where applicable, and review status metadata.
Unless otherwise noted, the dataset materials in this repository are licensed under the Creative Commons Attribution-NoDerivatives 4.0 International License (CC BY-ND 4.0). See LICENSE for details.… See the full description on the dataset page: https://huggingface.co/datasets/Anonymous-zxcvbnm/IO-Bench.Strudel-Synth
Strudel-Synth
Strudel-Synth is a synthetic corpus of 21,174 (MIDI, Strudel) pairs for
training and evaluating MIDI-to-Strudel decompilation, introduced in
Decomposer: Learning to Decompile Symbolic Music to Programs.
🎹 Live demo: anonymousgiraffe/decomposer-demo
🤗 Model: anonymousgiraffe/Decomposer-Qwen3-8B
Each pair consists of a Strudel program distilled from Claude-Opus-4.6 (conditioned on independently sampled musical and code-style seeds) and the MIDI produced by… See the full description on the dataset page: https://huggingface.co/datasets/anonymousgiraffe/Strudel-Synth.submission14717_fictionalqa_reformatted_triviaqa
Reformatted TriviaQA for use alongside FictionalQA
Repository: omitted
Paper: omitted
Dataset Description
This dataset is a simple derived view of the validation data from the original TriviaQA dataset hosted by the original creators at hf.co/datasets/mandarjoshi/trivia_qa. To create this view, we extract the wikipedia articles associated with each question, as well as a simplified answer list, and then we create a few versions of the resulting data for use as… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-aardvark/submission14717_fictionalqa_reformatted_triviaqa.