Back to Arena

Cognita

by TrueFoundry

System Card

OrganizationTrueFoundry
Released2024-01
Architecturevector-rag / Modular production RAG with incremental indexing
DetailsComponents for data sources, metadata (Postgres/Prisma), LLM gateway, vector DB (Qdrant/SingleStore), async indexing, API server. Multiple retrievers with similarity, query decomposition, reranking, incremental indexing.
Parameters—
Domainrag-retrieval
Open SourceYes
WebsiteVisit
CodeRepository
truefoundryincremental-indexingproductionarchived

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Long-Context Retrieval
2/5
RULER
664p
NIAH
no data
LooGLE
no data
LongBench
603p
∞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
62.913p
HotpotQA
63.626p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
63.64p
Sources:Cognita (truefoundry/cognita); evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Cognita (truefoundry/cognita); evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Cognita (truefoundry/cognita); evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Cognita (truefoundry/cognita); evaluated on RAGAS: Automated Evaluation of Retrieval-Augmented Generation (Exploding Gradients, 2309)Cognita (truefoundry/cognita); evaluated on RULER: What's the Real Context Size of Your Long-Context Language Models (NVIDIA, 2404)