Zhongyi Han

Contextual contamination in few-shot vision-language learning is mitigated by a novel semantic purification and reinforcement-learning gating mechanism, unlocking superior performance.

Within the tested range, concentration sets neither the destination nor the pace of collapse; the pace follows whose text fills the pool; what emerges instead is an invariance.

Tightening top-p, which cuts the low-probability tail at generation time, nearly stops collapse within three generations and stabilizes six checkpoints spanning the whole spectrum together, while data-side filtering slows collapse without stopping it.