Back to Arena

kNN-LM

by Stanford / Facebook AI Research (Khandelwal et al.)

System Card

OrganizationStanford / Facebook AI Research (Khandelwal et al.)
Released2019-11
Architectureexternal-memory-network / Linearly interpolated kNN over LM embedding space
DetailsInterpolates pretrained LM predictions with a kNN distribution over the LM's embedding space. Nearest neighbors drawn from any text collection enable domain adaptation without retraining.
Parameters—
Domainrag-retrievallifelong-learning
Open SourceYes
PaperView Paper
CodeRepository
iclr-2020nearest-neighbordatastorememorization

Capability Profile

Benchmark Scores

6 of 14 benchmarks
Long-Context Retrieval
1/5
RULER
no data
NIAH
no data
LooGLE
no data
LongBench
603p
∞Bench
no data
Multi-Turn Recall
2/2
LoCoMo
64.917p
MemoryBank
617p
Cross-Session Memory
1/1
LongMemEval
68.122p
Multi-Hop QA
2/3
BABILong
78.278p
MultiHop-RAG
no data
HotpotQA
69.542p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:kNN-LM paper (arXiv:1911.00172); evaluated on BABILong: Testing the Limits of LLMs with Long-Context Reasoning-in-a-Haystack (AIRI, 2406)kNN-LM paper (arXiv:1911.00172); evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)kNN-LM paper (arXiv:1911.00172); evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)kNN-LM paper (arXiv:1911.00172); evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)kNN-LM paper (arXiv:1911.00172); evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)kNN-LM paper (arXiv:1911.00172); evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)