File size: 12,847 Bytes
298c12a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
source,rows,license,url
amex,12911,cc-by-4.0,https://huggingface.co/datasets/Yuxiang007/AMEX
multimodal_mind2web,7613,openrail,https://huggingface.co/datasets/osunlp/Multimodal-Mind2Web
onejev/distill,6202,apache-2.0,https://github.com/OmniJev/OneJev
breakfast,5826,cc-by-4.0,https://serre.lab.brown.edu/breakfast-actions-dataset.html
typed-decisions-synth,4734,mit,https://huggingface.co/datasets/n4ze3m/typed-decisions-synth
osworld_verified_trajs,4634,mit,https://huggingface.co/datasets/xlangai/ubuntu_osworld_verified_trajs
agent_reward_bench,2226,research-only,https://huggingface.co/datasets/McGill-NLP/agent-reward-bench
videofeedback,2137,non-commercial (part),https://huggingface.co/datasets/TIGER-Lab/VideoFeedback
star,2136,apache-2.0 (labels); images: non-commercial,https://prior.allenai.org/projects/charades
misactbench,2128,cc-by-4.0,https://huggingface.co/datasets/osunlp/MisActBench
msrvtt,1731,none stated,https://www.microsoft.com/en-us/research/publication/msr-vtt-a-large-video-description-dataset-for-bridging-video-and-language/
mlfoundations_osworld_trajs,1588,none stated,https://huggingface.co/datasets/mlfoundations-cua-dev/osworld-trajectories
retention/onejev/distill,1513,apache-2.0,https://github.com/OmniJev/OneJev
diving48,1485,none stated,http://www.svcl.ucsd.edu/projects/resound/dataset.html
cua_speedrun_trajectories,1410,none stated,https://huggingface.co/datasets/anonymousmypcbench/cua-speedrun-trajectories
clevrer,1376,none stated,http://clevrer.csail.mit.edu/
nextqa,1362,mit (labels); images: non-commercial,https://github.com/doc-doc/NExT-QA
nebius__SWE-agent-trajectories,1316,cc-by-4.0,https://huggingface.co/datasets/nebius/SWE-agent-trajectories
konvid1k,1085,non-commercial,https://database.mmsp-kn.de/konvid-1k-database.html
dcagent-tb2-traces,1075,none stated,https://huggingface.co/DCAgent2
jevlite_dataset,1064,cc-by-sa-4.0,https://huggingface.co/datasets/vagmi/jevlite_dataset
dtap_bench,993,apache-2.0,https://huggingface.co/datasets/AI-Secure/DTap-Bench-Agent-Trajectories
swebench_experiments,966,none stated,https://github.com/swe-bench/experiments
captaincook4d,953,apache-2.0,https://captaincook4d.github.io/captain-cook/
aqa7,803,research-only,http://rtis.oit.unlv.edu/datasets.html
terminal-bench-2-leaderboard,799,apache-2.0,https://huggingface.co/datasets/harborframework/terminal-bench-2-leaderboard
gqa,762,cc-by-4.0,https://cs.stanford.edu/people/dorarad/gqa/download.html
osworld,670,mit,https://huggingface.co/datasets/xlangai/ubuntu_osworld_verified_trajs
imagerewarddb,659,apache-2.0,https://huggingface.co/datasets/zai-org/ImageRewardDB
s3_submissions_text,658,none stated,https://sierra-tau-bench-public.s3.us-west-2.amazonaws.com/
genai_arena,653,cc-by-4.0,https://huggingface.co/datasets/TIGER-Lab/GenAI-Bench
fgvc_aircraft,650,non-commercial,https://www.robots.ox.ac.uk/~vgg/data/fgvc-aircraft/
cauldron_clevr,620,cc-by-4.0,https://cs.stanford.edu/people/jcjohns/clevr/
workbench,566,mit,https://github.com/olly-styles/WorkBench
SWE-bench__SWE-smith-trajectories,556,mit,https://huggingface.co/datasets/SWE-bench/SWE-smith-trajectories
agiqa3k,529,mit,https://github.com/lcysyzxdxc/AGIQA-3k-Database
openhands_webarena,459,none stated,https://huggingface.co/datasets/OpenHandsCommunity/eval-output-webarena
snli_ve,443,non-commercial (images),https://github.com/necla-ml/SNLI-VE
flickr30k,442,non-commercial,https://shannon.cs.illinois.edu/DenotationGraph/
cauldron_tallyqa,441,apache-2.0 (labels); images: source terms,https://github.com/manoja328/TallyQA_dataset
a_okvqa,440,apache-2.0 (labels); images: source terms,https://github.com/allenai/aokvqa
vizwiz,440,cc-by-4.0,https://vizwiz.org/tasks-and-datasets/vqa/
cauldron_vqav2,438,cc-by-4.0 (labels); images: source terms,https://visualqa.org/terms.html
cauldron_nlvr2,438,cc-by-4.0 (labels); images: research-only,https://github.com/lil-lab/nlvr
scienceqa,437,cc-by-nc-sa-4.0,https://github.com/lupantech/ScienceQA
seerray_androidworld_eval,436,mit,https://huggingface.co/datasets/SeerRay-Lab/Android-World-Eval
nebius__SWE-rebench-openhands-trajectories,421,cc-by-4.0,https://huggingface.co/datasets/nebius/SWE-rebench-openhands-trajectories
cauldron_vsr,385,cc-by-4.0 (labels); images: source terms,https://github.com/cambridgeltl/visual-spatial-reasoning
cauldron_iconqa,382,cc-by-nc-sa-4.0,https://iconqa.github.io/
cauldron_visual7w,382,mit (labels); images: source terms,https://github.com/yukezhu/visual7w-toolkit
cauldron_figureqa,381,none stated,https://github.com/Maluuba/FigureQA
exgentic-agent-llm-traces-v2,349,none stated,https://huggingface.co/datasets/Exgentic/agent-llm-traces-v2
SWE-Gym__OpenHands-Sampled-Trajectories,348,none stated,https://huggingface.co/datasets/SWE-Gym/OpenHands-Sampled-Trajectories
cauldron_mapqa,343,cc-by-sa-4.0,https://github.com/OSU-slatelab/MapQA
agentsuite_tau2_bench,330,none stated,https://huggingface.co/datasets/AgentSuite/tau2-bench-trajectories
tau2_bench_github,320,mit,https://github.com/sierra-research/tau2-bench
hf_tarsur385_tau2_bench_simulations,318,none stated,https://huggingface.co/datasets/tarsur385/tau2-bench-simulations
cub200,315,non-commercial,https://www.vision.caltech.edu/datasets/cub_200_2011/
food101,315,research-only,https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-101/
oxford_pets,314,cc-by-sa-4.0 (labels); images: owners,https://www.robots.ox.ac.uk/~vgg/data/pets/
flowers102,312,none stated,https://www.robots.ox.ac.uk/~vgg/data/flowers/102/
eurosat,311,mit,https://github.com/phelber/EuroSAT
dtd,311,research-only,https://www.robots.ox.ac.uk/~vgg/data/dtd/
stanford_cars,310,none stated,https://ai.stanford.edu/~jkrause/cars/car_dataset.html
gtsrb,308,none stated,https://benchmark.ini.rub.de/
s3_tau_knowledge_paper,276,none stated,https://sierra-tau-bench-public.s3.us-west-2.amazonaws.com/
cauldron_ai2d,273,cc-by-sa-4.0,https://registry.opendata.aws/allenai-diagrams/
ava,272,none stated,https://github.com/imfing/ava_downloader
aadb,269,research-only,https://github.com/aimerykong/deepImageAestheticsAnalysis
tarsur385__swebench-verified-trajectories,251,mit,https://huggingface.co/datasets/tarsur385/swebench-verified-trajectories
cauldron_tqa,242,cc-by-nc-3.0,https://prior.allenai.org/projects/tqa
terminal-wrench,224,apache-2.0,https://huggingface.co/datasets/few-sh/terminal-wrench-trajectories
retention/DeveloperOats/DBPedia_Classes,207,cc0-1.0,https://huggingface.co/datasets/DeveloperOats/DBPedia_Classes
retention/clinc/clinc_oos,206,cc-by-3.0,https://huggingface.co/datasets/clinc/clinc_oos
agentdojo-trajectories,204,none stated,https://huggingface.co/datasets/sungjuncho/agentdojo-trajectories
kadid10k,195,pixabay,https://database.mmsp-kn.de/kadid-10k-database.html
SWE-Gym__OpenHands-Verifier-Trajectories,194,none stated,https://huggingface.co/datasets/SWE-Gym/OpenHands-Verifier-Trajectories
koniq10k,192,research-only,https://database.mmsp-kn.de/koniq-10k-database.html
flive,191,cc-by-nc-sa-4.0,https://github.com/baidut/PaQ-2-PiQ
hf_kermitco_qwen3_5_9b_tau2bench_telecom_traces,188,none stated,https://huggingface.co/datasets/KermitCO/qwen3.5-9B-tau2bench-telecom-traces
live_c,164,other (UT Austin notice),https://live.ece.utexas.edu/research/ChallengeDB/
retention/openbmb/UltraFeedback,162,mit,https://huggingface.co/datasets/openbmb/UltraFeedback
retention/prometheus-eval/Feedback-Collection,160,cc-by-4.0,https://huggingface.co/datasets/prometheus-eval/Feedback-Collection
tau_bench_v1_github,159,mit,https://github.com/sierra-research/tau-bench
retention/nvidia/HelpSteer2,153,cc-by-4.0,https://huggingface.co/datasets/nvidia/HelpSteer2
retention/legacy-datasets/banking77,142,cc-by-4.0,https://huggingface.co/datasets/legacy-datasets/banking77
theagentcompany-experiments,142,none stated,https://github.com/TheAgentCompany/experiments
openguardrails-tb21,137,mit,https://huggingface.co/openguardrails
hf_cloudsurf_software_tau2_trajectories_cloudsurf_4b_fc,132,cc-by-4.0,https://huggingface.co/datasets/cloudsurf-software/tau2-trajectories-cloudsurf-4b-fc
hf_debdootmiitd_tau3_bench_qwen3_6_35b_a3b_v0,132,apache-2.0,https://huggingface.co/datasets/debdootmiitd/tau3-bench-qwen3.6-35b-a3b-v0
R2E-Gym__R2EGym-Verifier-Trajectories,132,none stated,https://huggingface.co/datasets/R2E-Gym/R2EGym-Verifier-Trajectories
JetBrains-Research__agent-trajectories-swe-bench-test-minus-verified,130,mit,https://huggingface.co/datasets/JetBrains-Research/agent-trajectories-swe-bench-test-minus-verified
agentsuite_tau_bench,129,none stated,https://huggingface.co/datasets/AgentSuite/tau-bench-trajectories
hf_monte_inc_tau2_banking_archived,124,apache-2.0,https://huggingface.co/datasets/monte-inc/tau2-banking-archived
hf_lalitkumarps_sapientai_agent_opus_5_tau2_trajectories,123,mit,https://huggingface.co/datasets/lalitkumarps/sapientai-agent-opus-5-tau2-trajectories
webarena_v2_traces,120,none stated,https://github.com/web-arena-x/webarena/blob/main/resources/README.md
retention/google/civil_comments,116,cc0-1.0,https://huggingface.co/datasets/google/civil_comments
s3_tau_voice_paper_text,108,none stated,https://sierra-tau-bench-public.s3.us-west-2.amazonaws.com/
retention/tasksource/tasksource-instruct-v0,100,mixed,https://huggingface.co/datasets/tasksource/tasksource-instruct
openhands_miniwob,96,none stated,https://huggingface.co/datasets/OpenHandsCommunity/eval-output-miniwob
retention/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1,95,apache-2.0,https://huggingface.co/datasets/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1
retention/Muennighoff/natural-instructions,92,mixed,https://huggingface.co/datasets/Muennighoff/natural-instructions
retention/Arsive/toxicity_classification_jigsaw,86,apache-2.0,https://huggingface.co/datasets/Arsive/toxicity_classification_jigsaw
retention/nyu-mll/multi_nli,85,mixed,https://huggingface.co/datasets/nyu-mll/multi_nli
togethercomputer__CoderForge-Preview-32B-SWE-Bench-Verified-Evaluation-trajectories,81,none stated,https://huggingface.co/datasets/togethercomputer/CoderForge-Preview-32B-SWE-Bench-Verified-Evaluation-trajectories
retention/alisawuffles/WANLI,79,cc-by-4.0,https://huggingface.co/datasets/alisawuffles/WANLI
hf_monte_inc_tau2_banking_baselines,77,apache-2.0,https://huggingface.co/datasets/monte-inc/tau2-banking-baselines
hf_monte_inc_tau2_airline_baselines,74,none stated,https://huggingface.co/datasets/monte-inc/tau2-airline-baselines
hf_changdae_tau2_uq_artifacts,74,mit,https://huggingface.co/datasets/changdae/tau2-uq-artifacts
hf_monte_inc_tau2_banking_eval_results,72,apache-2.0,https://huggingface.co/datasets/monte-inc/tau2-banking-eval-results
hf_ziyiii0_0_trajectory2task_30b_a3b_taubench,70,none stated,https://huggingface.co/datasets/Ziyiii0-0/Trajectory2Task-30b-a3b-taubench
hf_florin_inc_tau2_retail_trajectories,67,apache-2.0,https://huggingface.co/datasets/florin-inc/tau2-retail-trajectories
hf_kargarisaac_tau2_retail_nablo_react_trajectories,63,mit,https://huggingface.co/datasets/kargarisaac/tau2-retail-nablo-react-trajectories
hf_xiaomingneu_tau3_qwen35_4b_retail_baseline,60,mit,https://huggingface.co/datasets/xiaomingneu/tau3-qwen35-4b-retail-baseline
hf_mindbomber_aana_tau2_bench_gpt41mini_1trial,53,mit,https://huggingface.co/datasets/mindbomber/aana-tau2-bench-gpt41mini-1trial
s3_voice_labels,45,none stated,https://sierra-tau-bench-public.s3.us-west-2.amazonaws.com/
snorkel_tau2_verified_airline_code_agents,42,apache-2.0,https://huggingface.co/datasets/snorkelai/Tau2-Bench-Verified-Airline-With-Code-Agents
hf_kermitco_qwen3_5_9b_tau2bench_retail_traces,32,none stated,https://huggingface.co/datasets/KermitCO/qwen3.5-9B-tau2bench-retail-traces
retention/google/boolq,31,cc-by-sa-3.0,https://huggingface.co/datasets/google/boolq
hf_kermitco_qwen3_5_9b_tau2bench_retail_baseline_traces,28,mit,https://huggingface.co/datasets/KermitCO/qwen3.5-9B-tau2bench-retail-baseline-traces
snorkel_tau2_airline_code_agents,27,apache-2.0,https://huggingface.co/datasets/snorkelai/Tau2-Bench-Airline-With-Code-Agents
AlienKevin__SWE-bench-Multilingual-trajectories,24,mit,https://huggingface.co/datasets/AlienKevin/SWE-bench-Multilingual-trajectories
hf_bojieli_pine_tau2_voice_gpt41_usersim,17,none stated,https://huggingface.co/datasets/bojieli/pine-tau2-voice-gpt41-usersim
hf_naclin_qwen3_5_9b_tau2_retail_dpo,10,none stated,https://huggingface.co/datasets/Naclin/qwen3.5-9b-tau2-retail-dpo
hf_kirutew17654321_tau2bench_gpt41_retail,8,mit,https://huggingface.co/datasets/kirutew17654321/tau2bench-gpt41-retail
hf_ai_sherpa_repro_tau2bench_bundle,8,none stated,https://huggingface.co/datasets/ai-sherpa/repro-tau2bench-bundle
hf_samtiz_tau2_voice_grok_tool_mentor_trajectories,4,none stated,https://huggingface.co/datasets/samtiz/tau2-voice-grok-tool-mentor-trajectories