List your company

Search-augmented multi-turn question answering on TriviaQA

64.6Exact Match (EM) AccuracySIOP
59.60860.90462.263.496May 6, 2026

Evaluation results

MethodLinks
SIOPBackbone=Qwen3-8B2026.0564.673.6
GRPO (gold-sup.)Backbone=Qwen3-8B2026.0564.372.8
IGPO (gold-sup.)Backbone=Qwen3-8B2026.0564.173
EMPOBackbone=Qwen3-8B2026.056371.7
GRPO (gold-sup.)Backbone=Qwen3-4B2026.0562.370.3
TTRLBackbone=Qwen3-8B2026.0562.372.2
EMPOBackbone=Qwen3-4B2026.0561.169.6
TTRLBackbone=Qwen3-4B2026.0560.769.1
SIOPBackbone=Qwen3-4B2026.0560.469.1
IGPO (gold-sup.)Backbone=Qwen3-4B2026.0559.868.2