datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
static-analysis-evalA dataset of 76 Python programs taken from real Python open source projects (top 100 on GitHub),
where each program is a file that has exactly 1 vulnerability as detected by a particular static analyzer (Semgrep), used in the paper Patched MOA: optimizing inference for diverse software development tasks.
OpenAI used the synth-vuln-fixes and fine-tuned
a new version of gpt-4o is now the SOTA on this benchmark. More details and code is available from their repo.
More details on the benchmark… See the full description on the dataset page: https://huggingface.co/datasets/patched-codes/static-analysis-eval.rm-static
Dataset Card for "rm-static"
Split of hh-static used for training reward models after supervised fine-tuning.
openarm_staticDiSH-Bench-static-v2
DiSH-Bench static v2 — test release
A simulated multi-receiver spatial speech benchmark, with aligned single-receiver
and multi-receiver questions over the same physical scenes. This release contains
122,256 test questions, 2,000 physical scenes, 4,400 original four-channel FOA
recordings, and 200 rooms. There are 58,829 single-input questions, 53,518
two-input questions and 9,909 three-input questions. Questions share recordings;
they are not independent acoustic scenes.
Source… See the full description on the dataset page: https://huggingface.co/datasets/Heiheihaha17/DiSH-Bench-static-v2.aera_semi_pnp_dr_16_06_2026_skip3_delta_no_go_home_no_static_smoothedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "AR4_MK3",
"total_episodes": 2855,
"total_frames": 399698,
"total_tasks": 156,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 23,
"splits": {
"train": "0:2855"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Purple69/aera_semi_pnp_dr_16_06_2026_skip3_delta_no_go_home_no_static_smoothed.aera_semi_pnp_dr_02_05_2026_skip3_delta_no_go_home_no_static_smoothed__0005This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "AR4_MK3",
"total_episodes": 5349,
"total_frames": 362999,
"total_tasks": 72,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 25,
"splits": {
"train": "0:5349"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Purple69/aera_semi_pnp_dr_02_05_2026_skip3_delta_no_go_home_no_static_smoothed__0005.bps-statictable-qrels-fullstatic-hhStatic split of Anthropic's Helpful Harmless dataset. Contains base-online and rejection sampled outputs.
Static-FineBenchqwen3-30b-0718-gpt55-static-balanced-7872-spare-games-envs
qwen3-30B-A3B-0718-gpt55-static-balanced-7872-400 — generated environments
Environments generated by the SPARE proposer during training run
h5wf1kei (qwen3-30B-A3B-0718-gpt55-static-balanced-7872-400), recovered from the spare-viz durable cache.
The run's scratch directory no longer exists; this dataset is the surviving copy.
Games
194
Steps covered
9 (step 0–115)
With recovered skill
0
With hint
0
Actor / proposer model… See the full description on the dataset page: https://huggingface.co/datasets/msr-spare-1/qwen3-30b-0718-gpt55-static-balanced-7872-spare-games-envs.wildfire-risk-static-datastatic-rlhf-interface-datapcbslm-static-v2-unsloth-vlm
PCBSLM static-v2 Unsloth VLM
Portable multimodal Unsloth dataset for PCB layout/document-grounded training.
The JSONL splits use Unsloth/Gemma-style chat messages:
{
"messages": [
{"role": "user", "content": [
{"type": "image", "image": "assets/raw_docs/.../images/page.png"},
{"type": "text", "text": "instruction..."}
]},
{"role": "assistant", "content": [
{"type": "text", "text": "{...json answer...}"}
]}
]
}
Files… See the full description on the dataset page: https://huggingface.co/datasets/henry1477/pcbslm-static-v2-unsloth-vlm.EverMemBench-Static
EverMemBench-S: Evaluating Evidence Access under Dense Semantic Interference
💻 Code: EverMind-AI/EverMemBench-Static
Overview
EverMemBench-S (EMB-S) is an adversarial Needle-in-a-Haystack benchmark built on a 326M-token MemoryBank with 160,280 documents across 8 domains. It evaluates long-context models and retrieval systems under dense semantic interference — where near-miss documents create realistic confusion that standard NIAH benchmarks cannot capture.
1,225… See the full description on the dataset page: https://huggingface.co/datasets/EverMind-AI/EverMemBench-Static.methods2test_small_static
methods2test_small (cleaned) + static method-level context
Static context for unit-test generation, computed with tree-sitter from each row's source only
(no repository checkout; target is never read).
Configs
default: train / validation / test from NDe2V/methods2test_small_cleaned @ 3d6a2f2061736202832b7ac2bb4e84e1c70668e8
methods2test_runnable: test from andstor/methods2test_runnable (fm+fc+c+m+f+t+tc) @ ed8cded666b2f7d101ba04a7a8d121b6249d8c39… See the full description on the dataset page: https://huggingface.co/datasets/NDe2V/methods2test_small_static.aera_semi_pnp_dr_16_06_2026_skip3_delta_no_go_home_no_static_smoothed_v2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "AR4_MK3",
"total_episodes": 2855,
"total_frames": 428406,
"total_tasks": 156,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 23,
"splits": {
"train": "0:2855"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Purple69/aera_semi_pnp_dr_16_06_2026_skip3_delta_no_go_home_no_static_smoothed_v2.spare-gpt55-static-corpus
SPARE GPT-5.5 Grounded Cognitive Multi-Turn Games
This public dataset contains 7,872 validated Python game environments for actor-only SPARE training.
Six cognitive skills, exactly 1,312 environments per skill
Generated with GPT-5.5 and grounded by spice_megascience_15k.jsonl
Grounding corpus SHA-256: a36a928b4940b5b5d9e3f4cb5804a94c69462360943adb3be14613c82f0f72c0
Maximum 25 turns and 32K generation context
Every environment passes load, reset, step, and replay validation with… See the full description on the dataset page: https://huggingface.co/datasets/msr-spare-1/spare-gpt55-static-corpus.component-static-buildsHEC3R-ckpt-real_abl_static_priordata_statics"Data Statics" data for PhilEO Bench
The Data Statics for the PhilEO Bench evaluation framework
Paper "Scaling-up the Pretraining of the Earth Observation Foundation Model PhilEO to the MajorTOM Dataset" - ESA Φ-lab
http://github.com/ESA-PhiLab/PhilEO-MajorTOM
Paper: http://arxiv.org/pdf/2506.14765
Also: Paper: http://arxiv.org/pdf/2506.14765v1
Scaling-Up the Pretraining of the Earth Observation Foundation Model PhilEO to the MajorTOM Dataset
Also: PhilEO Bench:… See the full description on the dataset page: https://huggingface.co/datasets/NikolaosDionelis2023/data_statics.aera_semi_pnp_dr_02_05_2026_skip3_delta_no_go_home_no_static_smoothedbenji-lora-datasetbps-statictable-qrels-allusc-vecs-v1-chunks-v1-s4096-o512-sentence-transformers-static-retrieval-mrl-en-v1usc-vecs-v1-chunks-v1-s2048-o256-sentence-transformers-static-retrieval-mrl-en-v1query-hard-pos-neg-doc-pairs-statictablecode2lora-static
Code2LoRA-Static — static track (single snapshot)
The Static track of RepoPeftBench exactly as consumed by the paper's
main-results table (Code2LoRA-Static, single anchor snapshot per
repository). Contains only the paper-used QnAs (post quality-filter). For
the out-of-distribution slice see code2lora/repopeftbench-ood.
config
split
rows
qna
train
39,612
qna
cr_val
6,213
qna
cr_test
6,414
qna
ir_val
4,833
qna
ir_test
5,222
repos
train / cr_val / cr_test
409 /… See the full description on the dataset page: https://huggingface.co/datasets/code2lora/code2lora-static.methods2test_small_static
methods2test_small (cleaned) + static method-level context
Static context for unit-test generation, computed with tree-sitter from each row's source only
(no repository checkout; target is never read).
Configs
default: train / validation / test from NDe2V/methods2test_small_cleaned @ 3d6a2f2061736202832b7ac2bb4e84e1c70668e8
methods2test_runnable: test from andstor/methods2test_runnable (fm+fc+c+m+f+t+tc) @ ed8cded666b2f7d101ba04a7a8d121b6249d8c39… See the full description on the dataset page: https://huggingface.co/datasets/quill1112/methods2test_small_static.usc-vecs-v1-chunks-v1-s8192-o512-sentence-transformers-static-retrieval-mrl-en-v1PickPlace_staticMarker
PickPlace_staticMarker (TsFile)
Apache TsFile version of BlankHead/PickPlace_staticMarker.
Overview
A LeRobot robot manipulation dataset. Each frame holds the commanded action and observed observation.state joint positions; camera views are stored as videos in the original dataset.
Episodes: 76
Frames: 26911
Sampling rate: 30 fps
Tasks: 1
Split: a single train split
Robot: so_follower
Cameras (not uploaded): wrist, up, side
Schema (TsFile structure)… See the full description on the dataset page: https://huggingface.co/datasets/THULab/PickPlace_staticMarker.