-
oaimli/pgpo_grpo_full_scitrek_qwen3_4b_instruct_2507
Text Generation • 4B • Updated • 474 -
oaimli/pgpo_grpo_full_scitrek_gemma3_4b_it
Image-Text-to-Text • 4B • Updated • 100 -
oaimli/pgpo_grpo_proxy_scitrek_qwen3_4b_instruct_2507
Text Generation • 4B • Updated • 257 -
oaimli/pgpo_grpo_proxy_scitrek_gemma3_4b_it
Image-Text-to-Text • 4B • Updated • 88
Miao Li
oaimli
AI & ML interests
Natural Language Processing
Recent Activity
updated a collection 9 days ago
SciTrek updated a collection 9 days ago
SciTrek updated a model 9 days ago
oaimli/scitrek_grpo_full_longrlvr_qwen3_4b_instruct_2507Organizations
None yet
ProxyCoT
Models for Long-Context Reasoning Through Proxy-Based Chain-of-Thought Tuning (ACL 2026)
-
oaimli/longtune_scitrek_reasoning_reinforcement_qwen
Text Generation • 4B • Updated • 4 -
oaimli/longtune_scitrek_grounding_reinforcement_qwen_5_300
Text Generation • 4B • Updated • 5 -
oaimli/longtune_scitrek_grounding_reinforcement_qwen_0_300
Text Generation • 4B • Updated • 7 -
oaimli/longtune_scitrek_reasoning_reinforcement_gemma
Image-Text-to-Text • 4B • Updated • 4
Others
LongPT
Long-context post-training
SciTrek
Models for the paper of "Who Gets Cited Most? Benchmarking Long-Context Language Models on Scientific Articles"
-
oaimli/scitrek_qwen25_7b_instruct_1m_sft_input_length
Text Generation • 8B • Updated • 147 -
oaimli/scitrek_qwen25_7b_instruct_1m_sft_question_focus
Text Generation • 8B • Updated • 3 -
oaimli/scitrek_qwen25_7b_instruct_1m_sft_sql_type
Text Generation • 8B • Updated • 9 -
oaimli/scitrek_qwen25_7b_instruct_1m_grpo_curriculum_192
Text Generation • 8B • Updated • 136
PGPO
-
oaimli/pgpo_grpo_full_scitrek_qwen3_4b_instruct_2507
Text Generation • 4B • Updated • 474 -
oaimli/pgpo_grpo_full_scitrek_gemma3_4b_it
Image-Text-to-Text • 4B • Updated • 100 -
oaimli/pgpo_grpo_proxy_scitrek_qwen3_4b_instruct_2507
Text Generation • 4B • Updated • 257 -
oaimli/pgpo_grpo_proxy_scitrek_gemma3_4b_it
Image-Text-to-Text • 4B • Updated • 88
LongPT
Long-context post-training
ProxyCoT
Models for Long-Context Reasoning Through Proxy-Based Chain-of-Thought Tuning (ACL 2026)
-
oaimli/longtune_scitrek_reasoning_reinforcement_qwen
Text Generation • 4B • Updated • 4 -
oaimli/longtune_scitrek_grounding_reinforcement_qwen_5_300
Text Generation • 4B • Updated • 5 -
oaimli/longtune_scitrek_grounding_reinforcement_qwen_0_300
Text Generation • 4B • Updated • 7 -
oaimli/longtune_scitrek_reasoning_reinforcement_gemma
Image-Text-to-Text • 4B • Updated • 4
SciTrek
Models for the paper of "Who Gets Cited Most? Benchmarking Long-Context Language Models on Scientific Articles"
-
oaimli/scitrek_qwen25_7b_instruct_1m_sft_input_length
Text Generation • 8B • Updated • 147 -
oaimli/scitrek_qwen25_7b_instruct_1m_sft_question_focus
Text Generation • 8B • Updated • 3 -
oaimli/scitrek_qwen25_7b_instruct_1m_sft_sql_type
Text Generation • 8B • Updated • 9 -
oaimli/scitrek_qwen25_7b_instruct_1m_grpo_curriculum_192
Text Generation • 8B • Updated • 136
Others