Instruction Tuning Datasets - a skymizer Collection

skymizer 's Collections

Instruction Tuning Datasets

Pre-Training Datasets

Domain-Specific Datasets

Alignment Algorithm Papers

Other Awesome datasets

Instruction Tuning Datasets

updated Jun 18

For both SFT and DPO

tatsu-lab/alpaca

Viewer • Updated May 22, 2023 • 52k • 24.4k • 701

Note SFT
elichen3051/alpaca52k-alignment-handbook

Viewer • Updated Jun 7 • 52k • 42 • 1
yahma/alpaca-cleaned

Viewer • Updated Apr 10, 2023 • 51.8k • 20.2k • 587

Note SFT
HuggingFaceH4/ultrachat_200k

Viewer • Updated 27 days ago • 515k • 12.6k • 474

Note SFT
openbmb/UltraFeedback

Viewer • Updated Dec 29, 2023 • 64k • 1.55k • 335

Note SFT
HuggingFaceH4/ultrafeedback_binarized

Viewer • Updated 27 days ago • 187k • 5.69k • 239

Note SFT; DPO
HuggingFaceH4/deita-10k-v0-sft

Viewer • Updated Mar 1 • 20k • 233 • 31

Note SFT
HuggingFaceH4/orca_dpo_pairs

Viewer • Updated Apr 14 • 12.9k • 160 • 25

Note DPO
HuggingFaceH4/cai-conversation-harmless

Viewer • Updated Feb 2 • 44.8k • 198 • 13

Note SFT; DPO
shahules786/orca-chat

Viewer • Updated Jul 25, 2023 • 198k • 163 • 110

Note SFT
lmsys/lmsys-chat-1m

Viewer • Updated Jul 27 • 1M • 62.3k • 595

Note SFT; DPO
What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning

Paper • 2312.15685 • Published Dec 25, 2023 • 17
HuggingFaceH4/no_robots

Viewer • Updated Apr 18 • 10k • 1.49k • 447

Note SFT
HuggingFaceH4/SystemChat

Viewer • Updated Mar 12 • 7.02k • 56 • 9
HuggingFaceH4/deita-6k-v0-sft

Viewer • Updated Jan 9 • 12k • 44
Open-Orca/OpenOrca

Viewer • Updated Oct 21, 2023 • 2.91M • 10.7k • 1.34k
Open-Orca/slimorca-deduped-cleaned-corrected

Viewer • Updated May 9 • 182k • 748 • 15
allenai/WildChat-1M

Viewer • Updated 26 days ago • 838k • 2.29k • 283
mlabonne/orpo-dpo-mix-40k

Viewer • Updated 26 days ago • 44.2k • 1.5k • 243
argilla/ultrafeedback-binarized-preferences-cleaned

Viewer • Updated Dec 11, 2023 • 60.9k • 7.28k • 124
argilla/distilabel-capybara-dpo-7k-binarized

Viewer • Updated Jul 16 • 7.56k • 845 • 174
argilla/OpenHermesPreferences

Viewer • Updated Mar 1 • 989k • 2.6k • 199

Note DPO
OpenAssistant/oasst2

Viewer • Updated Jan 11 • 135k • 1.43k • 213