arxiv:2508.07616
Aswin Ravikumar Rangsasamy Veerasamy
rrvaswin
·
AI & ML interests
Transformers, SSMs
Recent Activity
submitted a paper 5 days ago
Mid-Training with Self-Generated Data Improves Reinforcement Learning in Language Models updated a model 21 days ago
rrvaswin/qwen_STaR_RLOrganizations
models 117
rrvaswin/qwen_STaR_RL
8B • Updated • 246
rrvaswin/qwen_4b_RL
8B • Updated • 244
rrvaswin/qwen_Vanilla_RL
8B • Updated • 13
rrvaswin/qwen_8b_RL
8B • Updated • 17
rrvaswin/qwen_16b_RL
8B • Updated • 15
rrvaswin/qwen_32b_RL
8B • Updated • 13
rrvaswin/qwen_star_baseline
8B • Updated • 247
rrvaswin/qwen_32b_distill_baseline
8B • Updated • 39
rrvaswin/qwen_32b_SFT
8B • Updated • 47
rrvaswin/qwen_16b_SFT
8B • Updated • 148
datasets 30
rrvaswin/QwQ_64b_Distill
Viewer • Updated • 320k • 35
rrvaswin/qwen2.5_star
Viewer • Updated • 7.43k • 31
rrvaswin/65b_rewarded_top16
Viewer • Updated • 114k • 8
rrvaswin/qwq32b_distill_data_1
Viewer • Updated • 114k • 23
rrvaswin/qwq32b_distill_data
Viewer • Updated • 114k • 24
rrvaswin/qwen_2.5_behaviors_5k
Viewer • Updated • 325k • 269
rrvaswin/qwen_2.5_behaviors
Viewer • Updated • 464k • 1.11k
rrvaswin/analysis_1_to_16
Viewer • Updated • 7.41k • 5
rrvaswin/analysis_1_to_1
Viewer • Updated • 7.41k • 7
rrvaswin/rewarded_16b_gsm8k_incorrect
Viewer • Updated • 77.5k • 3