List your company

Question Answering on SQuAD Short-form QA (F1, AIL, RCO)

74.79F1 ScoreBatch (Full)
0.679619.919839.1658.4002May 30, 2026

Evaluation results

MethodLinks
Batch (Full)Backbone=Qwen2.5-3B-Instruct, Processing Mode=Offline2026.0574.7977.55100
Proactive-EntrBackbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming2026.0573.4853.1475
Proactive-AttnBackbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming2026.0571.6959.1789
Batch (Full)Backbone=Qwen-3-4B, Processing Mode=Offline2026.0570.7576.53100
Proactive-EntrBackbone=Qwen-3-4B, Processing Mode=Streaming2026.0562.4849.4171
Proactive-AttnBackbone=Qwen-3-4B, Processing Mode=Streaming2026.0560.1347.1869
Wait-9Backbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming, Wait-k interval=92026.0515.14-21.3219
Wait-7Backbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming, Wait-k interval=72026.0511.76-22.8915
Wait-7Backbone=Qwen-3-4B, Processing Mode=Streaming, Wait-k interval=72026.059.83-16.8130
Wait-5Backbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming, Wait-k interval=52026.058.91-25.2111
Wait-3Backbone=Qwen2.5-3B-Instruct, Processing Mode=Streaming, Wait-k interval=32026.058.03-27.4414
Wait-9Backbone=Qwen-3-4B, Processing Mode=Streaming, Wait-k interval=92026.054.661.0348
Wait-5Backbone=Qwen-3-4B, Processing Mode=Streaming, Wait-k interval=52026.053.74-3.4645
Wait-3Backbone=Qwen-3-4B, Processing Mode=Streaming, Wait-k interval=32026.053.53-5.1344