Tag
The paper introduces ARC, a training recipe for fairer relative advantage comparison in open-ended real-world interaction by conditioning rollouts on strategy, and presents INTER3, a paradigm for responsive user-agent interaction that reduces latency.
This paper presents a fair-comparison study of variational quantum circuits in diffusion models, introducing a squeeze-and-excitation scaffold to isolate quantum contributions. It finds functional parity with classical controls and identifies angle-embedding failures in score-based settings, offering a rigorous methodology and mechanistic analysis.
This paper demonstrates that global calibration metrics like Expected Calibration Error are confounded by model accuracy, and proposes ACE, an accuracy-controlled evaluation framework for fair comparison of large language models.