datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
RLVE-Qwen3-1.7B-Pass1-Rollouts
RLVE teacher rollouts — Qwen3-1.7B (pass@1)
Teacher rollouts for on-policy distillation on the RLVE environment suite.
Teacher / sampler: Qwen3-1.7B
Source prompts: RLVE train split — 9000 questions across RLVE-Eval Gym
environments (counting / combinatorics / optimization tasks)
Sampling: 1 sample/question (pass@1) = 9000 records,
temperature 0.7, max 4096 new tokens
Rewards: inline RLVE-Eval Gym verifier score (continuous, in [-1, 1]).
Teacher accuracy (reward>0): 20 / 9000 =… See the full description on the dataset page: https://huggingface.co/datasets/CL-From-Nothing/RLVE-Qwen3-1.7B-Pass1-Rollouts.RLVE_envs65_data1000rlve_teacherrl-verifier-pitfalls_hacking_dataThis is the hacking dataset associated with the paper "Pitfalls of Rule- and Model-based Verifiers -- A Case Study on Mathematical Reasoning." In this paper, we design and release a "Hacking Dataset" of 13 adversarial patterns (e.g., gibberish, HTML tags, empty symbols) to evaluate verifier robustness.
rlve_recoverablerlve-eval-rolloutsrlve_teacher_topk16_20K
RLVE Teacher Top-16 Logit Data (20K)
Teacher top-k logit sidecar data for continuation-style KD-SFT warmup
(see compute_teacher_topk_logprobs.py / KDContinuationDataset).
Each row holds, per response token, the teacher's top-16 (+ forced true token)
candidate token ids and their log-probabilities, joined to the base dataset by
row_id.
Configs
rlve_offline_20K — 20,000 rows (rlve_offline_20K_teacher_top16.parquet)
rlve_rose_20K — 20,000 rows… See the full description on the dataset page: https://huggingface.co/datasets/CL-From-Nothing/rlve_teacher_topk16_20K.rlve_rose_20Krlve_offline_20KRLVE-Qwen3-4B-Thinking-2507-Pass8-Rollouts
RLVE teacher rollouts — Qwen3-4B-Thinking-2507 (pass@8)
Teacher rollouts for on-policy distillation on the RLVE environment suite.
Teacher / sampler: Qwen3-4B-Thinking-2507
Source prompts: RLVE train split — 9000 questions across 18 environments
(counting / combinatorics / optimization tasks)
Sampling: 8 samples/question (pass@8) = 72000 records,
temperature 1.0 (sample.sh default 0.7 -> here T per run), max 16384 new tokens
Rewards: recomputed offline with the RLVE-Eval Gym… See the full description on the dataset page: https://huggingface.co/datasets/CL-From-Nothing/RLVE-Qwen3-4B-Thinking-2507-Pass8-Rollouts.rlve_offline_20K_POPE_prefix_pass1_qwen3-1.7b
RLVE offline-20K POPE-prefix completions — Qwen3-1.7B (pass1)
Prefix-conditioned completions generated by Qwen3-1.7B over the
rlve_offline_20K_POPE_prefix prompt set (20000 records, 1 sample/prompt).
Produced by SLURM job 6580578 (vLLM, tp=2), 2026-06-15.
Fields
index, sample_id, prompt, prefix, response, answer, rewards
⚠️ Caveat on rewards
The inline rewards field is all 0.0 — this is the known inline-Gym-verifier
artifact (same as the old… See the full description on the dataset page: https://huggingface.co/datasets/CL-From-Nothing/rlve_offline_20K_POPE_prefix_pass1_qwen3-1.7b.rlve-32b-k1024-sft
RLVE × Qwen3-32B, K=1024 采样(SFT 数据)
Qwen3-32B 在 RLVE 训练集 9000 道题上采样, 每题 2 条, 每条上限 1024 token。
共 18000 条。
采样设置
与本项目所有 RLVE 评测逐项一致, 以避免训练/评测的 prompt 错配:
prompt: tokenizer.apply_chat_template(msgs, add_generation_prompt=True),
不传 enable_thinking(即 Qwen3 默认开启 thinking)
temperature 0.7, top_p 0.9, top_k -1, n=2, max_tokens 1024
使用前必读:这批数据全部是截断的
含 </think>
0 / 18000 (0.0%)
含 \boxed{}
0 / 18000 (0.0%)
response 字符数
中位 3818 / 均值 3708 / 最大… See the full description on the dataset page: https://huggingface.co/datasets/SeanWang0027/rlve-32b-k1024-sft.roofline_sft_RLVE_Qwen3_4B_Thinking_2507mixed_sft_RLVE_Qwen3_4B_Thinking_2507RLVE-Test-Qwen3-1.7B-GRPO-step70-Pass8
RLVE test eval — GRPO step70 (pass@8)
Evaluation rollouts on the RLVE test split.
Model: grpo_train_Qwen3-1.7B-SFT-rlve-20K-1epoch (GRPO, step 70)
Source prompts: RLVE test split — 180 questions (RLVE-Eval Gym environments)
Sampling: 8 samples/question (pass@8) = 1440 records, temperature 0.7,
max 16384 new tokens
Rewards: inline RLVE-Eval Gym verifier score (continuous, in [-1, 1]).
Record-level accuracy (reward>0): 216 / 1440 = 15.0%, mean reward -0.677
pass@8 (>=1 of 8… See the full description on the dataset page: https://huggingface.co/datasets/CL-From-Nothing/RLVE-Test-Qwen3-1.7B-GRPO-step70-Pass8.ptdbench-rlve-hyper-task-009-dataset
PTDBench dataset snapshot: task_009
This repository stores the immutable runtime dataset snapshot for one
materialized PTDBench task. It intentionally excludes model weights and
training checkpoints.
PTDBench family: rlve_hyper
Source evaluation metric: eval/HELD-OUT_ENVIRONMENTS_128
Provenance: RLVE repository snapshot under its MIT license; bundled upstream benchmark notices remain applicable.
License: MIT
The artifact manifest records every hydrated runtime path, byte size… See the full description on the dataset page: https://huggingface.co/datasets/LIF1014/ptdbench-rlve-hyper-task-009-dataset.rl-verified-v1a small effort to create a math dataset for RL :)
RLVE-d5_Qwen3-1.7B_with-continuationptdbench-rlve-hyper-task-001-dataset
PTDBench dataset snapshot: task_001
This repository stores the immutable runtime dataset snapshot for one
materialized PTDBench task. It intentionally excludes model weights and
training checkpoints.
PTDBench family: rlve_hyper
Source evaluation metric: eval/HELD-OUT_ENVIRONMENTS_128
Provenance: RLVE repository snapshot under its MIT license; bundled upstream benchmark notices remain applicable.
License: MIT
The artifact manifest records every hydrated runtime path, byte size… See the full description on the dataset page: https://huggingface.co/datasets/LIF1014/ptdbench-rlve-hyper-task-001-dataset.ptdbench-rlve-hyper-task-008-dataset
PTDBench dataset snapshot: task_008
This repository stores the immutable runtime dataset snapshot for one
materialized PTDBench task. It intentionally excludes model weights and
training checkpoints.
PTDBench family: rlve_hyper
Source evaluation metric: eval/HELD-OUT_ENVIRONMENTS_128
Provenance: RLVE repository snapshot under its MIT license; bundled upstream benchmark notices remain applicable.
License: MIT
The artifact manifest records every hydrated runtime path, byte size… See the full description on the dataset page: https://huggingface.co/datasets/LIF1014/ptdbench-rlve-hyper-task-008-dataset.rlve-rollouts-nemotron-cascade-8b-qwen3-1.7bRLVE-d5_Qwen3-4B-Thinking-2507_env73rlve-multitask-qwen3-4b-rollouts-n4-tokens16384rlve-student-prefix-qwen3-4b-thinking-qwen3-1.7bptdbench-rlve-hyper-task-012-dataset
PTDBench dataset snapshot: task_012
This repository stores the immutable runtime dataset snapshot for one
materialized PTDBench task. It intentionally excludes model weights and
training checkpoints.
PTDBench family: rlve_hyper
Source evaluation metric: eval/HELD-OUT_ENVIRONMENTS_128
Provenance: RLVE repository snapshot under its MIT license; bundled upstream benchmark notices remain applicable.
License: MIT
The artifact manifest records every hydrated runtime path, byte size… See the full description on the dataset page: https://huggingface.co/datasets/LIF1014/ptdbench-rlve-hyper-task-012-dataset.RLVE-d5_Qwen3-1.7B_env73rlve-teacher-completion-qwen3-4b-thinkingrlve_hardptdbench-rlve-hyper-task-016-dataset
PTDBench dataset snapshot: task_016
This repository stores the immutable runtime dataset snapshot for one
materialized PTDBench task. It intentionally excludes model weights and
training checkpoints.
PTDBench family: rlve_hyper
Source evaluation metric: eval/HELD-OUT_ENVIRONMENTS_128
Provenance: RLVE repository snapshot under its MIT license; bundled upstream benchmark notices remain applicable.
License: MIT
The artifact manifest records every hydrated runtime path, byte size… See the full description on the dataset page: https://huggingface.co/datasets/LIF1014/ptdbench-rlve-hyper-task-016-dataset.rl-verifiedA small effort to create a dataset for RL :)