List your company

Summarization on LongBench GovReport (ROUGE-L)

34.61ROUGE-LBaseline
16.451621.165825.8830.5942Feb 13, 2026Feb 15, 2026Feb 17, 2026Feb 19, 2026Feb 21, 2026Feb 23, 2026Feb 26, 2026

Evaluation results

MethodLinks
BaselineModel=Llama 3.1-8B2026.0234.61
KIVIModel=Llama 3.1-8B2026.0234.47
KIVISinkModel=Llama 3.1-8B2026.0234.33
InnerQBaseModel=Llama 3.1-8B2026.0234.33
BaselineModel=Llama 3.2-3B2026.0233.86
InnerQSmallModel=Llama 3.1-8B2026.0233.8
KIVISinkModel=Llama 3.2-3B2026.0233.76
InnerQHybridModel=Llama 3.1-8B2026.0233.62
InnerQBaseModel=Llama 3.2-3B2026.0233.38
InnerQHybridModel=Llama 3.2-3B2026.0232.73
InnerQSmallModel=Llama 3.2-3B2026.0232.25
KIVIModel=Llama 3.2-3B2026.0232.11
BaselineModel=Llama 3.2-1B2026.0229.27
InnerQBaseModel=Llama 3.2-1B2026.0228.89
KIVISinkModel=Llama 3.2-1B2026.0228.87
KIVIModel=Llama 3.2-1B2026.0228.57
BaselineModel=Llama 2-13B2026.0227.82
InnerQBaseModel=Llama 2-13B2026.0227.29
KIVISinkModel=Llama 2-13B2026.0227.1
KIVIModel=Llama 2-13B2026.0227.01
BaselineModel=Llama 2-7B2026.0226.88
InnerQHybridModel=Llama 3.2-1B2026.0226.72
InnerQSmallModel=Llama 3.2-1B2026.0226.7
KIVIModel=Llama 2-7B2026.0226.68
KIVISinkModel=Llama 2-7B2026.0226.64
InnerQHybridModel=Llama 2-13B2026.0226.38
InnerQBaseModel=Llama 2-7B2026.0226.12
InnerQSmallModel=Llama 2-13B2026.0225.69
InnerQHybridModel=Llama 2-7B2026.0225.18
CurvPruneContext Budget (B)=20482026.0221
RandomContext Budget (B)=20482026.0220.6
BM25+TexContext Budget (B)=20482026.0220.5
Head+TailContext Budget (B)=20482026.0220.4
BM25Context Budget (B)=20482026.0220.3
LLMLinguaContext Budget (B)=20482026.0219.8
Sel. CtxContext Budget (B)=20482026.0219.7
RecencyContext Budget (B)=20482026.0219.1
InnerQSmallModel=Llama 2-7B2026.0217.15