Paradot

by WithFeeling.AI

System Card

OrganizationWithFeeling.AI

Released2023-07

Architecturehierarchical-summary / Memory-to-Understanding Model (M2U)

DetailsParadot's proprietary Memory-to-Understanding Model is a matrix of LLMs ranging from several billion to tens of billions of parameters that captures a real-time evolving memory stream during conversations. Memories are categorized into levels based on their consolidation potential — from fleeting feelings to core facts — building a long-term evolving understanding of each user. Users can directly view, edit, delete, and favorite memories in a visible memory stream, making it the most user-transparent memory system among companion apps.

Parameters—

Domainpersonalizationepisodic-sessionlifelong-learning

Open SourceNo

WebsiteVisit

companionmemory-to-understandingemotional-aivisible-memoryuser-editableevolving

Capability Profile

Benchmark Scores

6 of 14 benchmarks

Long-Context Retrieval

0/5

RULER

no data

NIAH

no data

LooGLE

no data

LongBench

no data

∞Bench

no data

Multi-Turn Recall

2/2

LoCoMo

76.463p

MemoryBank

76.958p

Cross-Session Memory

1/1

LongMemEval

74.743p

Multi-Hop QA

1/3

BABILong

76.864p

MultiHop-RAG

no data

HotpotQA

no data

Agent Task Memory

1/1

AgentBench-Mem

7226p

Personalization

1/1

PerLTQA

76.645p

Factuality / Grounding

0/1

RAGAS

no data

Sources:Paradot vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Paradot vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Paradot vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)Paradot vendor documentation; evaluated on PerLTQA: A Personal Long-Term Memory Question Answering Dataset (PolyU, 2402)Paradot vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)Paradot vendor documentation; evaluated on BABILong: Testing the Limits of LLMs with Long-Context Reasoning-in-a-Haystack (AIRI, 2406)