List your company

Multi-turn Calibration on NQ

4.68ECE@1MTCal
2.942814.668926.39538.1211Apr 7, 2026

Evaluation results

MethodLinks
MTCalBackbone=Gemma2-9B-it2026.044.682.454.1222.714.82
MTCalBackbone=Llama3.1-8B-Instruct2026.045.186.076.0322.785.15
MTCalBackbone=Qwen2.5-7B-Instruct2026.046.223.44.721.023.71
DCalBackbone=Llama3.1-8B-Instruct2026.046.911.897.6123.587.65
DCalBackbone=Gemma2-9B-it2026.0410.4611.5814.5923.8513.98
DCalBackbone=Qwen2.5-7B-Instruct2026.0413.5517.5914.7122.1114.62
PSBackbone=Gemma2-9B-it2026.0418.3921.5518.9325.2217.15
SCBackbone=Gemma2-9B-it2026.0420.4321.6728.9332.7822.83
PSBackbone=Llama3.1-8B-Instruct2026.0420.6533.8125.2729.4324.22
SCBackbone=Llama3.1-8B-Instruct2026.0420.9421.7922.8228.6417.29
SLBackbone=Gemma2-9B-it2026.0422.4326.1520.5227.6718.76
SLBackbone=Llama3.1-8B-Instruct2026.0424.1234.9727.5431.3124.41
SCBackbone=Qwen2.5-7B-Instruct2026.0425.0328.3526.1630.4119.03
PSBackbone=Qwen2.5-7B-Instruct2026.0432.0240.0932.6532.9229.63
P(True)Backbone=Llama3.1-8B-Instruct2026.0432.7543.5341.0540.2534.46
VerbalBackbone=Qwen2.5-7B-Instruct2026.0433.9445.7543.4943.9139.91
SLBackbone=Qwen2.5-7B-Instruct2026.0434.1149.5143.9642.0531.93
P(True)Backbone=Qwen2.5-7B-Instruct2026.0440.5456.346.4645.2235.83
VerbalBackbone=Llama3.1-8B-Instruct2026.0441.6751.1243.1541.5736.28
P(True)Backbone=Gemma2-9B-it2026.0444.3743.6338.2137.531.22
VerbalBackbone=Gemma2-9B-it2026.0448.1147.3940.7440.5837.38