Reinforce-Ada: An Adaptive Sampling Framework for Reinforce-Style LLM Training
Paper • 2510.04996 • Published • 16
Reinforce-Ada
Training & test sets and finetuned models
-
-
weqweasdas/math500
Viewer • Updated • 500 • 81 -
weqweasdas/aime_hmmt_brumo_cmimc_amc23
Viewer • Updated • 230 • 41 -
weqweasdas/olympiadbench
Viewer • Updated • 675 • 289 -
weqweasdas/minerva_math
Viewer • Updated • 272 • 33
weqweasdas/from_default_filtered_openr1
Viewer • Updated • 75k • 50Note Prompt set used for data processing
RLHFlow/reinforce_ada_hard_prompt
Viewer • Updated • 15.7k • 55 • 2Note Selected hard prompts used to train Qwen2.5-Math-7B and Qwen3-4B-Instruct-2507
RLHFlow/reinforce_ada_easy_prompt
Viewer • Updated • 24.3k • 36Note Selected easy prompts used to train Qwen2.5-Math-7B
RLHFlow/reinforce_ada_hard_prompt_llama
Viewer • Updated • 15k • 26Note Selected hard prompts used to train Llama-3.2-3B-Instruct
RLHFlow/Llama-3.2-3B-Instruct-Reinforce-Ada-balance-hard
4B • Updated • 4Note Checkpoint from step=400 and trained on the hard prompt set
RLHFlow/Qwen2.5-Math-7B-Reinforce-Ada-balance-hard
8B • Updated • 4Note Checkpoint from step=400 and trained on the hard prompt set.
RLHFlow/Qwen3-4B-Instruct-2507-Reinforce-Ada-balance-hard
4B • Updated • 5 • 1Note Checkpoint from step=400 and trained on the hard prompt set.
RLHFlow/Qwen2.5-Math-7B-Reinforce-Ada-balance-easy
8B • Updated • 7Note Checkpoint from step=500 and trained on the easy prompt set.
RLHFlow/reinforce_ada_simple_prompt_1-5b
Viewer • Updated • 25k • 52Note Selected easy prompts used to train Qwen2.5-Math-1.5B
RLHFlow/reinforce_ada_hard_prompt_1-5b
Viewer • Updated • 13.3k • 49Note Selected hard prompts used to train Qwen2.5-Math-1.5B
-
RLHFlow/Qwen2.5-Math-1-5B-Reinforce-Ada-balance-easy
2B • Updated • 9 -
RLHFlow/Qwen2.5-Math-1-5B-Reinforce-Ada-balance-hard
Updated • 5 -
RLHFlow/Qwen2.5-Math-1.5B-DAPO-easy
2B • Updated • 4 -
RLHFlow/Qwen2.5-Math-1.5B-GRPO-n8-easy
2B • Updated • 11