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

LLM Inference Efficiency on Short sequence prompts

0.2TTFT (s)

FedRAG

-0.38123.54197.46511.3881May 25, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2026.05
0.233.77173.1475
2026.05
0.2220.674675
2026.05
0.2536.15156.4775
2026.05
0.2547.04141.0175
2026.05
0.8516.38452.59683
2026.05
0.8618.18439.03683
2026.05
0.9319.68466.64558
2026.05
0.9813.25379.52683
2026.05
1.3910.19220.66867
2026.05
1.649.12357.671,427
2026.05
3.1317.79941.4484
2026.05
3.2311.7301.171,147
2026.05
3.525.941,181.292,618
2026.05
3.659.91328.011,287
2026.05
4.8630.971,526.88580
2026.05
5.225.661,418.02628
2026.05
5.24.861,677.83,878
2026.05
6.225.131,791.253,458
2026.05
9.064.072,406.174,298
2026.05
9.0924.452,117.09676
2026.05
9.681.794,224.678,163
2026.05
11.931.624,721.7310,851
2026.05
12.971.425,135.9212,195
2026.05
14.731.295,620.9113,539