Share your thoughts, 1 month free Claude Pro on usSee more

Retrieval-Augmented Question Answering on DeepSearch PopQA

64Success Rate

ActiveContext

Updated 5d ago

Evaluation Results

Method	Links
ActiveContext 2026.04		64	7.4
Full Context 2026.04		62	36.4
Synapse 2026.04		54	100.5
Synapse 2026.04		48	76.4
Training-Free GRPO 2026.04		44	47.2
ActiveContext 2026.04		40.9	7.3
Synapse 2026.04		40	47.8
Gemini-3.0-flash 2026.04		40	8
Synapse 2026.04		38	50.5
ContextCurator 2026.04		38	7.6
Training-Free GRPO 2026.04		36	31
Full Context 2026.04		34	33.8
AWM 2026.04		34	34.8
gpt-4o 2026.04		34	12
Full Context 2026.04		32	35.1
Full Context 2026.04		32	35.1
Mem0 2026.04		32	47.9
Training-Free GRPO 2026.04		32	36.2
gpt-4o-mini 2026.04		32	10.8
Gemini-2.5-flash 2026.04		30	8.2
Qwen2.5-7B-Instruct 2026.04		28	9.9
Training-Free GRPO 2026.04		26	75.9
ActiveContext 2026.04		26	7.7