Independent project. Not a U.S. government website.Independent project. Not a United States government website.
Support our projectSupport usSupport our project · Buy me a coffee (opens Buy Me a Coffee)[paper] [demo samples] [Original repository] Introduction The Expresso dataset is a high-quality (48kHz) expressive speech dataset that includes both expressively render…
100 likes3,593 downloads in the last 30 daysCC BY-NC 4.0
🤗 [LongAlign Dataset] • 💻 [Github Repo] • 📃 [LongAlign Paper] LongAlign is the first full recipe for LLM alignment on long context. We propose the LongAlign-10k datas…
100 likes2,784 downloads in the last 30 daysQuestion answeringNo license given
BigCodeBench-Complete: Code Completion based on the structured docstrings. BigCodeBench-Instruct: Code Generation based on the NL-oriented instructions. The overal…
94 likes167,769 downloads in the last 30 daysApache 2.0
Stack-Edu is a 125B token dataset of educational code filtered from The Stack v2, precisely the curated training corpus of StarCoder2 models denoted StarCoder2Data. It i…
84 likes15,885 downloads in the last 30 daysNo license given
Zyda is a 1.3T language modeling dataset created by collecting open and high quality datasets and combining them and performing a uniform filtering and deduplication ste…
84 likes10,883 downloads in the last 30 daysText generationODC-By
The dataset consists of 4 versions: bigcode/the-stack-v2: the full "The Stack v2" dataset bigcode/the-stack-v2-dedup: based on the bigcode/the-stack-v2 but further near-…
70 likes1,041 downloads in the last 30 daysText generationOther licenseGated
This pretraining dataset, for generative AI model training, preserves high-value math and code while enriching it with diverse multilingual Q&A, fueling the next generat…
70 likes956 downloads in the last 30 daysOther license
Note that this collection licensed under ODC-BY. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. The Dolci Instruct…
63 likes3,488 downloads in the last 30 daysOtherODC-By
⚠️ WARNING: This dataset is intended ONLY for reproducing Olmo 3 7B ⚠️ For all other training use cases, including training from scratch, please utilize our primary dolm…
56 likes34,414 downloads in the last 30 daysText generationODC-By
This is the Dolma 3 pool, pre–quality upsampling and mixing. If you are interested in the data used to train Olmo 3 7B and Olmo 3 32B, visit allenai/dolma3_mix-6T-1025.…
42 likes20,909 downloads in the last 30 daysText generationODC-By
The Dolma 3 Mix (6T) is the collection of data used during the pretraining stage to train the Olmo-3-1125-32B model. This dataset is made up of ~6 trillion tokens from a…
39 likes47,113 downloads in the last 30 daysText generationODC-By
Sources include a mixture of existing reasoning traces: OpenThoughts 3 (Apache 2.0): Extended to 32K context length and downsampled code prompts to 16X multiple, to 941,…
32 likes3,008 downloads in the last 30 daysODC-By
This dataset contains the Meta evaluation result details for Llama-3.1-8B. The dataset has been created from 12 evaluation tasks. These tasks are triviaqa__wiki, mmlu_pr…
28 likes49 downloads in the last 30 daysLlama 3.1 licenseGated
Dolci-Think-RL is a deliberate reasoning RL dataset used for training Olmo-3-32B-Think model. It contains 102,026 high-quality prompts covering: Math Code Precise Instru…
27 likes1,449 downloads in the last 30 daysNo license given
Dolma 3 Dolmino dataset pool for Olmo 3 stage 2 annealing training This dataset contains the high-quality pool of data considered for the second stage of Olmo 3 32B. Dat…
27 likes26,907 downloads in the last 30 daysODC-By
The Dolma 3 Longmino Mix (100B) is the mixture of data used for the third stage of training for Olmo 3 32B model. Dataset Sources Source Type LC-s2pdf-REX 32k-64k Synth…
19 likes3,802 downloads in the last 30 daysODC-By
19 likes1,010 downloads in the last 30 daysNo license given
This dataset is licensed under ODC-BY. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. The Dolci Instruct DPO mixtur…
17 likes1,710 downloads in the last 30 daysODC-By
QA benchmark (test set) for the ether0 reasoning language model: This benchmark is made from commonly used tasks - like reaction prediction in USPTO/ORD, molecular capti…
17 likes3,690 downloads in the last 30 daysQuestion answeringCC BY 4.0
Dolci-Instruct-RL is the reinforcement learning dataset used to train the Olmo-3-7B-Instruct model.It contains 169,964 prompts spanning: Math Code Precise Instruction Fo…
15 likes526 downloads in the last 30 daysNo license given
Molmo2-Cap is a dataset of very long (average 900 words), detailed video captions. It can be used to pre-train and fine-tune vision-language models. Molmo2-Cap is part o…
15 likes497 downloads in the last 30 daysODC-By
This dataset contains 1024 60 second video clips of Bleeding Edge gameplay (75GB). The data has already been processed into the following format: 300x180 videos sampled…
12 likes814 downloads in the last 30 daysNo license given
This dataset is licensed under ODC-BY. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. The Dolci Think DPO mixture w…
11 likes276 downloads in the last 30 daysODC-By
The Dolma 3 Dolmino Mix (100B) is the mixture of high-quality data used for the second stage of training for Olmo 3 7B model. Dataset Sources Source Category Tokens Docu…
11 likes12,235 downloads in the last 30 daysText generationODC-By
The Dolma 3 Longmino Mix (50B) is the mixture of data used for the third stage of training for Olmo 3 7B model. Dataset Sources Source Type Tokens Docs LC-s2pdf-REX 32k-…
11 likes3,906 downloads in the last 30 daysODC-By
IMPORTANT: Please help us protect the integrity of this benchmark by not publicly sharing, re-uploading, or distributing the dataset. Humanity's Last Exam: Diamond 🌐 We…
10 likes856 downloads in the last 30 daysMITGated
This is the Dolma 3.5 pool. It contains no quality upsampling or mixing. This is an updated version of the Dolma 3 pool with additional quality filtering and more data s…
9 likes23,933 downloads in the last 30 daysText generationODC-By
Snapshot: 2026-09-28. 150 cases and 6,874 question instances. Default reference: consensus. Labels are model-generated references. Data Load configuration cases, questio…
7 likes1,732 downloads in the last 30 daysApache 2.0
5 likes4,740 downloads in the last 30 daysNo license given
hermes-4-70b-reasoning-40k Evaluation Results Summary Benchmark Score Metric Samples Overlong rate aime24 0.735 math_pass@1:64_samples 64 8.4% aime25 0.674 math_pass@1:6…
5 likes212 downloads in the last 30 daysNo license given
0.1%…
5 likes212 downloads in the last 30 daysNo license given
Snapshot: 2026-09-28. 111 cases and 1,124 question instances. Default reference: consensus. Labels are model-generated references. Data Load configuration cases, questio…
3 likes1,391 downloads in the last 30 daysNo license given
Snapshot: 2026-09-28. 204 cases and 3,287 question instances. Default reference: consensus. Labels are model-generated references. Data Load configuration cases, questio…
3 likes1,565 downloads in the last 30 daysNo license given
Snapshot: 2026-09-28. 240 cases and 1,820 question instances. Default reference: consensus. Labels are model-generated references. Data Load configuration cases, questio…
3 likes1,405 downloads in the last 30 daysNo license given
1 likes130 downloads in the last 30 daysNo license given
The Pile: Validation and Test Splits This repo contains the validation and test splits of The Pile, an 825 GiB English text dataset designed for training large language…
0 likes503 downloads in the last 30 daysText generationMIT
Help keep USASI useful.
Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).
About supporting this project