Across three independent inference seeds with the trained checkpoint and evaluation settings fixed, JAM’s sample standard deviation is below 1.5 points on all four benchmarks (range 0.31–1.34). Variability is lower on LongMemEval and LoCoMo than on NarrativeQA and HotpotQA.

Across three independent inference seeds with the trained checkpoint and evaluation settings fixed, JAM’s sample standard deviation is below 1.5 points on all four benchmarks (range 0.31–1.34). Variability is lower on LongMemEval and LoCoMo than on NarrativeQA and HotpotQA. · CiteArk

Measurement results

0/20 assessed

0 supported0 disputed or conflicting20 not assessed

The complete result group stays with its claim. Assessed rows come first; their judgments do not replace the whole-claim conclusion. Expand a row for conditions, sources, and evidence history.

Each row is a measurement with its reported value, latest available observation, and cumulative evidence state.
Metric and conditions	Reported	Latest observation	Evidence judgment
benchmark: LoCoMo F1 · statisticOrRun: Run 1	52.09 F1 points (0–100)	—	Not assessed
benchmark: LoCoMo F1 · statisticOrRun: Run 2	52.68 F1 points (0–100)	—	Not assessed
benchmark: LoCoMo F1 · statisticOrRun: Run 3	51.1 F1 points (0–100)	—	Not assessed
benchmark: LoCoMo F1 · statisticOrRun: Mean	51.96 F1 points (0–100)	—	Not assessed
benchmark: LoCoMo F1 · statisticOrRun: Std.	0.8 F1 points (0–100)	—	Not assessed
benchmark: LongMemEval Acc. · statisticOrRun: Run 1	65.2%	—	Not assessed
benchmark: LongMemEval Acc. · statisticOrRun: Run 2	64.6%	—	Not assessed
benchmark: LongMemEval Acc. · statisticOrRun: Run 3	65%	—	Not assessed
benchmark: LongMemEval Acc. · statisticOrRun: Mean	64.93%	—	Not assessed
benchmark: LongMemEval Acc. · statisticOrRun: Std.	0.31 percentage points	—	Not assessed
benchmark: NarrativeQA F1 · statisticOrRun: Run 1	45.72 F1 points (0–100)	—	Not assessed
benchmark: NarrativeQA F1 · statisticOrRun: Run 2	43.37 F1 points (0–100)	—	Not assessed

Across three independent inference seeds with the trained checkpoint and evaluation settings fixed, JAM’s sample standard deviation is below 1.5 points on all four benchmarks (range 0.31–1.34). Variability is lower on LongMemEval and LoCoMo than on NarrativeQA and HotpotQA.

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results