0.0%
Overall Score
0/0
Tests Passed
0
Total Bottles
0
v4.0 Extracted
Temporal Ordering at 1K Events
Track 1,000 timestamped events and answer "What happened BEFORE X?"
temporal
medium
Pass: 85%
Belief Evolution Over 500 Sessions
Track how a belief changed across 500 sessions, identify current version
temporal
hard
Pass: 80%
Multi-Hop Reasoning (Conversational: 3-5 hops)
Follow reasoning chains in conversational data (target: 3-5 hops, technical docs: 10+)
reasoning
medium
Pass: 75%
Entity Graph at 5K Nodes
Build and traverse a knowledge graph with 5,000 interconnected entities
graph
hard
Pass: 85%
Cross-Session Continuity (100 Sessions)
Validate memory retention across 100 session boundaries
continuity
medium
Pass: 90%
Scale Stress Test (1.5M Tokens)
Process and query a 1.5M token conversation (LongMemEval maximum)
scale
extreme
Pass: 70%
Causality Chain Tracking (1K Events)
Track cause→effect relationships in 1,000 event sequences
temporal
hard
Pass: 80%
Contradiction Resolution (Conversational: 10-50)
Detect contradictory beliefs in conversational data (realistic: 10-50 per 100 sessions)
reasoning
medium
Pass: 50%