List your company

Long-context Language Modeling on LongBench (Generation Score, Attention Recall)

50.6Generation ScoreRaBitQCache
1.87614.525527.17539.8245Jun 30, 2026

Evaluation results

MethodLinks
RaBitQCacheBudget=p = 0.95 (17.33%), Model=LLaMA-3.1-8B2026.0650.690.7
FullBudget=-(100%), Model=LLaMA-3.1-8B2026.0650.58100
Double SparseBudget=8192 (64.60%), Model=LLaMA-3.1-8B2026.0650.4796.9
QuestBudget=8192 (64.57%), Model=LLaMA-3.1-8B2026.0650.4397
Double SparseBudget=4096 (40.33%), Model=LLaMA-3.1-8B2026.0650.490.7
OracleBudget=1024 (11.4%), Model=LLaMA-3.1-8B2026.0650.3186.9
Double SparseBudget=1024 (11.42%), Model=LLaMA-3.1-8B2026.0650.2868.4
SparQBudget=Ratio=0.25 (25.00%), Model=LLaMA-3.1-8B2026.0650.1589.8
QuestBudget=4096 (40.29%), Model=LLaMA-3.1-8B2026.0649.2492.9
Double SparseBudget=256 (2.86%), Model=LLaMA-3.1-8B2026.0649.0642.7
QuestBudget=1024 (11.38%), Model=LLaMA-3.1-8B2026.0646.5282.6
QuestBudget=256 (2.84%), Model=LLaMA-3.1-8B2026.0636.7153.3
QuestBudget=4096 (40.29%), Model=Longchat-7B-v1.5-32k2026.0636.2886.1
RaBitQCacheBudget=p = 0.95 (17.33%), Model=Longchat-7B-v1.5-32k2026.0636.2189.8
QuestBudget=8192 (64.57%), Model=Longchat-7B-v1.5-32k2026.0636.1194.7
SparQBudget=Ratio=0.25 (25.00%), Model=Longchat-7B-v1.5-32k2026.0635.9193.1
OracleBudget=1024 (11.4%), Model=Longchat-7B-v1.5-32k2026.0635.8489.3
FullBudget=-(100%), Model=Longchat-7B-v1.5-32k2026.0635.78100
QuestBudget=1024 (11.38%), Model=Longchat-7B-v1.5-32k2026.0635.358.1
Double SparseBudget=8192 (64.60%), Model=Longchat-7B-v1.5-32k2026.0632.5867.4
QuestBudget=256 (2.84%), Model=Longchat-7B-v1.5-32k2026.0630.527.3
Double SparseBudget=4096 (40.33%), Model=Longchat-7B-v1.5-32k2026.0621.2847.8
Double SparseBudget=1024 (11.42%), Model=Longchat-7B-v1.5-32k2026.069.5220.4
Double SparseBudget=256 (2.86%), Model=Longchat-7B-v1.5-32k2026.063.7510