one-pass
Datasets
All datasets matching “one-pass”easyr1-103k-4MP-stage-one-temp-1_1-RL-min-2-pass-max-7-passone_pass_ingestionpassword-locked-dataset-one-tokenone-pass-sv-forms-synthetic
One-Pass SV-Forms synthetic corpus (Swedish)
This dataset is entirely synthetic. It was generated by a script from a concept catalogue, not
collected from people or customer submissions. Names and organisations are constructed,
email addresses use .invalid, and identifier-shaped values are generated locally.
They are not checked against registries; coincidental matches with real names or
identifiers cannot be ruled out. The dataset is published so the recipe behind… See the full description on the dataset page: https://huggingface.co/datasets/precisit/one-pass-sv-forms-synthetic.tmax-15k-filtered-pass-at-least-one-time-7k
tmax15k_full_qwen35_9b_pass_at_least_once_snapshot_20260916_185018
This snapshot contains 6922 unique Harbor tasks for which at least one
recorded trial had reward=1 in the source Qwen3.5-9B k=3 run.
Snapshot time (UTC): 2026-09-16T11:15:09.256006+00:00
Source task set: /mnt/sidi/OPD/data/tmax15k_harbor
Source run: /mnt/sidi/OPD/run_outputs/tmax15k_full_qwen35_9b_k3_conc16_20260914_162523
Source run complete at snapshot time: False
Completed/planned trials: 31023/43803… See the full description on the dataset page: https://huggingface.co/datasets/StephYang/tmax-15k-filtered-pass-at-least-one-time-7k.dapo_easy_one_third_sorted_by_pass_rate
DAPO easy subset, chosen by model pass rate
This dataset is created by modifying this dataset: DAPO. We thank the authors of the original dataset for releasing a nice dataset that enabled our research.
The main change from the original dataset:
We want a subset of the original dataset that is substantially easier, where a small LLM would retain good performance.
Therefore, we construct an easy subset of DAPO following the steps below:
We generate 32 responses per prompt on the… See the full description on the dataset page: https://huggingface.co/datasets/ftajwar/dapo_easy_one_third_sorted_by_pass_rate.
pr_ir_onepassrepro-full-batch-gradient-descent-outperforms-one-pass-sgd-sample-complexityrepro-full-batch-gd-outperforms-one-pass-sgd-sample-complexity-separationrepro-full-batch-gradient-descent-outperforms-one-pass-sgd-sample-complexity-separation-in-singlrepro-full-batch-gd-one-pass-sgd-single-index-nativerepro-full-batch-gradient-descent-outperforms-one-pass-sgd-sample-complexity-separation-in-s