Kepler operationalizes an auditable evidence contract: append-only evidence, complete-history retrodiction, conditional prediction checks, and mechanical replay.
Auditable world-model harness achieves perfect ARC-AGI-3 score
Kepler uses retrospective transition checks and conditional prediction checks to validate agent hypotheses, reaching 100.00 RHAE on all 25 public games without per-game model selection or score-conditioned reruns.
Independent
Wensen Wu
Independent Researcher
Research Digest··3 min read
The authors present Kepler, an open-source harness for ARC-AGI-3 that represents agent hypotheses as executable world models and validates them through retrodiction and prediction checks.
Why this paper
From Independent Researcher
In one line
Kepler’s auditable executable world models achieved 100.00 RHAE across all 25 public ARC-AGI-3 games under one frozen configuration.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (3 noted)
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§