Back to Arena

RETRO

by DeepMind (Borgeaud et al.)

System Card

OrganizationDeepMind (Borgeaud et al.)
Released2021-12
Architecturevector-rag / Chunked cross-attention over 2T-token BERT-indexed datastore
DetailsConditions an autoregressive LM on document chunks retrieved by local similarity to preceding tokens. Uses a frozen BERT retriever, differentiable encoder, and chunked cross-attention to attend over 2T tokens.
Parameters—
Domainrag-retrieval
Open SourceNo
PaperView Paper
WebsiteVisit
CodeRepository
icml-20222-trillion-tokenscross-attentionretrofit

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Data Transparency:5 estimated
Long-Context Retrieval
2/5
RULER
68.930pEstimated
NIAH
no data
LooGLE
no data
LongBench
603pEstimated
∞Bench
no data
Multi-Turn Recall
0/2
LoCoMo
no data
MemoryBank
no data
Cross-Session Memory
0/1
LongMemEval
no data
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
63.818pEstimated
HotpotQA
62.924pEstimated
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
65.728pEstimated
Sources:Arena estimate — derived from capability profile, not independently verified