Track 3 · #1Overall #4
docket
Awarded to Sergiu
OSS-engine (OpenCypher)
The full 500-question LongMemEval at 0.3511 against a 0.000 no-memory baseline.
Judges’ citation
Only entry that ran the FULL 500-question LongMemEval (0.3511 vs 0.000 no-memory baseline); rigorous refusal taxonomy; public per-question inspector; found real HydraDB cursor bug via 207-probe capability map.
Track 3 · Agent memory — LongMemEval, BEAM, MedLoCoMo