datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
total-300-random-jh-epoch4
total-300-random-jh-epoch4
Portable process-evaluation output. metadata.json is the lightweight source
for aggregate results; the JSONL files are directly loadable; and
artifacts.tar.gz losslessly preserves the original run directory.
Reasoning score: 0.3890625
Action score: 0.440625
Valid samples: 320/320
vsr_random
VSR: Visual Spatial Reasoning
This is the random set of VSR: Visual Spatial Reasoning (TACL 2023) [paper].
Usage
from datasets import load_dataset
data_files = {"train": "train.jsonl", "dev": "dev.jsonl", "test": "test.jsonl"}
dataset = load_dataset("cambridgeltl/vsr_random", data_files=data_files)
Note that the image files still need to be downloaded separately. See data/ for details.
Go to our github repo for more introductions.
Citation
If you find VSR… See the full description on the dataset page: https://huggingface.co/datasets/cambridgeltl/vsr_random.vertebrate-v1-issue473-fullwindow-cds-random-val
marin-dna/vertebrate-v1-issue473-fullwindow-cds-random-val
CDS full-window vertebrate projection sequences for the issue #473 random
validation control. The source is the immutable issue #417 accepted-sequence
table.
The split uniformly samples 16,384 original-orientation CDS rows
without replacement using seed 42. Sampling occurs before
reverse-complement augmentation. Selected rows are removed from training;
reverse complements are then added only to the remaining training… See the full description on the dataset page: https://huggingface.co/datasets/marin-dna/vertebrate-v1-issue473-fullwindow-cds-random-val.synthetic-social-networks
Synthetic Social Networks (Dataset)
Raw experimental outputs from the Synthetic Social Networks study:
59,776 in-character LLM-agent posts from 528 production trials, and
64,562 posts total when the original pipeline-verification runs are
included. The artifact combines an exploratory stage with a separately
frozen, preregistered 448-trial matched-exposure confirmation. Each production
trial includes peer-vote traces from in-character voting by other agents.… See the full description on the dataset page: https://huggingface.co/datasets/ranausmans/synthetic-social-networks.open-bottleneck-ranklong27b-slurm-364982-rollouts
Open Bottleneck RankLong 27B — Slurm array 364982
Compact rollout evidence archived from completed Slurm array 364982.
Config
Files / steps
Records
JSONL bytes
Note
rank_a40
60 (1–60)
15,360
66,445,670
Complete local rollout evidence
rank_a80
54 (1–54)
13,824
59,386,451
Includes the cancelled arm's final dumped step (54.jsonl)
Each JSONL record contains input, output, gts, score, acc,
response_length, grouprel_reward, and step.
Only rollout evidence is archived… See the full description on the dataset page: https://huggingface.co/datasets/ryankim17920/open-bottleneck-ranklong27b-slurm-364982-rollouts.ransomware-leak-site-victims
Ransomware leak-site victims
Every victim listing ThreatCluster has collected first-hand from ransomware and extortion leak sites: the group, the named organisation, when it appeared, and the sector and country where known.
Built from the ThreatCluster corpus. 20,627 rows, snapshot generated 2026-09-06.
Fields
Field
Description
group_name
Ransomware or extortion group that published the listing
victim_name
Organisation named by the group
country… See the full description on the dataset page: https://huggingface.co/datasets/threatcluster/ransomware-leak-site-victims.TRIM-gpt-oss-120b-separate-neighbors-only-para-random-feature-num
TRIM Agent Reasoning Messages (HF Public Export)
This directory is a Hugging Face-friendly public export of the TRIM agent reasoning SFT data.
What Is Included
Provider: vllm
Model: gpt-oss-120b
SFT mode: local_neighbor_only
Splits present: train
Records in this export manifest: 10056
Tasks in this split: AMES, BBB_Martins, Bioavailability_Ma, CYP2C9_Substrate_CarbonMangels, CYP2D6_Substrate_CarbonMangels, CYP3A4_Substrate_CarbonMangels, Carcinogens_Lagunin, ClinTox… See the full description on the dataset page: https://huggingface.co/datasets/Kiria-Nozan/TRIM-gpt-oss-120b-separate-neighbors-only-para-random-feature-num.TVR_rankingRolePlay_Collection_random_ShareGPT
RolePlay_Collection_random_ShareGPT
This collection of random datasets for roleplay has been gathered from various sources. It has been processed, cleaned, and grammar-checked. However, it still requires significant work to be fully usable. Additional cleaning is necessary. I hope this proves helpful to someone.
repro-understanding-lora-as-knowledge-memory-an-empirical-analysis-traces
Agent traces
Agent sessions published from a Trackio Logbook.
repro-randomized-feasibility-traces
Agent traces
Agent sessions published from a Trackio Logbook.
randomizer-5Mэто случайные цифорки
Нехнаю зачем я это создал
lm-eval-EleutherAI_tampered-deep-ignorance-random-init-fp-adversarial-20251104_051748
Dataset Card for Evaluation run of EleutherAI/tampered-deep-ignorance-random-init-fp-adversarial-20251104_051748
Dataset automatically created during the evaluation run of model EleutherAI/tampered-deep-ignorance-random-init-fp-adversarial-20251104_051748
The dataset is composed of 2 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 11 run(s). Each run can be found as a specific split in each configuration, the split being… See the full description on the dataset page: https://huggingface.co/datasets/EleutherAI/lm-eval-EleutherAI_tampered-deep-ignorance-random-init-fp-adversarial-20251104_051748.turkish-question-departmentmusical-instruments-ranker-unseen-resampled-20260924-v1
Musical Instruments original ranker protocol: resampled v1
A training-only expansion of the original ranking dataset. Same 1,648 training queries, same original negative-generation rules. Four distinct orderings of each complete reference bundle give 6,592 positive rows. Two independent negatives per strategy are requested: random set, role collision, wrong item, wrong query. There are 13,150 negatives/comparisons (exactly twice the original 6,575): random set 3,296; role… See the full description on the dataset page: https://huggingface.co/datasets/flavianv/musical-instruments-ranker-unseen-resampled-20260924-v1.tell-human-detectors
TELL Human Detectors Split
This dataset is a prompt-disjoint validation/test split of human_detectors.json from
Jenna Russell, Marzena Karpinska, and Mohit Iyyer, "People who frequently use
ChatGPT for writing tasks are accurate and robust detectors of AI-generated text"
(arXiv:2501.15654).
The split is intended as a research artifact for evaluating AI-vs-human writing
detectors and explanation-quality methods. It preserves the upstream fields, with
one documented issue type… See the full description on the dataset page: https://huggingface.co/datasets/suraj-ranganath/tell-human-detectors.gpn_combined_random_uniform_10Mb_256random-user-profiles
Random User Profiles
2,000 synthetic user profiles with random ids, ages, plans and scores. For testing only.
Note: This is randomly generated synthetic data with no real-world meaning. Generated for testing and demonstration purposes.
pacman_eval_easy_braidthin_random100_v1_20260820
Pacman Easy braided/thinned — random 100 v1
This is the exact 100-episode Pacman Easy shard used by the published UWM trajectories.
Source dataset: pacman_2d_easy_braidthin_oracle_success_v1_20260815 (500 episodes)
Selection: MT19937 random sample without replacement (seed 2026082001)
Records: 100
JSONL SHA-256: 788d71aeb295e3a54ddfca870d8c7c720f59185c43ff21a3105cdc6ecf8d52b0
Full model trajectories:… See the full description on the dataset page: https://huggingface.co/datasets/novastar111/pacman_eval_easy_braidthin_random100_v1_20260820.random_effect_examplespeed-ranking
Speed Ranking
All 31 working dispatchAI models ranked by CPU inference speed.
🚀 dispatchAI
pacman_eval_forward_stress_g12_random100_20260818
Forward stress G12 — deterministic random 100
This is the exact 100-episode Pacman evaluation shard used for the final UWM trajectories.
Source dataset: pacman_2d_easy_g12_f2_greedyfood_trap_forward_eval_v1_20260818
Sampling: Python MT19937 without replacement, seed 2026081812
Records: 100
JSONL SHA-256: 8118f7b275d345a1a6558df4f53c6de3382b23d8c663ebd112660e52219a6051
Trajectories: https://huggingface.co/datasets/novastar111/uwm_paper_eval_trajectories
sample_manifest.json… See the full description on the dataset page: https://huggingface.co/datasets/novastar111/pacman_eval_forward_stress_g12_random100_20260818.mosim-humanoid-walk-random-ar1-dt005
MoSim Humanoid Walk Random AR(1), dt=0.005
Humanoid Walk state transition dataset collected from DMC with random AR(1)-style actions.
Format
Each .npz contains:
data: float32 array with shape (num_trajectories, 1000, state_dim + action_dim + state_dim).
metadata: object with state_dim, action_dim, and DMC qpos/qvel sizes when available.
Each transition row is [s_t, a_t, s_(t+1)].
For Humanoid Walk:
state_dim = 55
action_dim = 21
row dimension = 55 + 21 + 55 = 131… See the full description on the dataset page: https://huggingface.co/datasets/Weyl09/mosim-humanoid-walk-random-ar1-dt005.Nexesenex__Llama_3.2_1b_RandomLego_RP_R1_0.1-details
Dataset Card for Evaluation run of Nexesenex/Llama_3.2_1b_RandomLego_RP_R1_0.1
Dataset automatically created during the evaluation run of model Nexesenex/Llama_3.2_1b_RandomLego_RP_R1_0.1
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/Nexesenex__Llama_3.2_1b_RandomLego_RP_R1_0.1-details.pacman_eval_two_ghost_braidthin_random100_20260817
HF Two-Ghost braided/thinned — deterministic random 100
This is the exact 100-episode Pacman evaluation shard used for the final UWM trajectories.
Source dataset: pacman_2d_easy_g9f8_two_ghost_braidthin_oracle_success_test_v1_20260817
Sampling: Python MT19937 without replacement, seed 2026081702
Records: 100
JSONL SHA-256: fbaf50a9a9799384b2cc1daf5c4c5c7e81c3fed8a23b01cc3b5b53747f3ed034
Trajectories: https://huggingface.co/datasets/novastar111/uwm_paper_eval_trajectories… See the full description on the dataset page: https://huggingface.co/datasets/novastar111/pacman_eval_two_ghost_braidthin_random100_20260817.lm-eval-EleutherAI_deep-ignorance-random-init
Dataset Card for Evaluation run of EleutherAI/deep-ignorance-random-init
Dataset automatically created during the evaluation run of model EleutherAI/deep-ignorance-random-init
The dataset is composed of 0 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the… See the full description on the dataset page: https://huggingface.co/datasets/EleutherAI/lm-eval-EleutherAI_deep-ignorance-random-init.pacman_eval_forward_stress_g8_random100_20260818
Forward stress G8 — deterministic random 100
This is the exact 100-episode Pacman evaluation shard used for the final UWM trajectories.
Source dataset: pacman_2d_easy_g8_f2_greedyfood_trap_forward_eval_v1_20260818
Sampling: Python MT19937 without replacement, seed 2026081808
Records: 100
JSONL SHA-256: 5044700e1e317981e889a34936593e8d95f431fa34e0e95705b1cc758168a4c6
Trajectories: https://huggingface.co/datasets/novastar111/uwm_paper_eval_trajectories
sample_manifest.json… See the full description on the dataset page: https://huggingface.co/datasets/novastar111/pacman_eval_forward_stress_g8_random100_20260818.autotrain-data-amdal-mining-llama2-7b-clean5k-dataset-geogpt-fineweb-randomvisgym_pacman_2d_random
VisGym Pacman2D Random Policy World-Model Trajectories
This dataset contains random-policy interaction trajectories for the custom VisGym Pacman2D environment.
Each row is one environment trajectory. The history entries include image_prev, image, and
image_next base64 JPEG frames, the prompt shown at the step, the sampled action, reward, and
environment info. The policy samples move actions uniformly and samples the valid stop action with
the configured random stop probability after… See the full description on the dataset page: https://huggingface.co/datasets/novastar112/visgym_pacman_2d_random.