manifold

This paper introduces MAG, a manifold-guided framework for semi-supervised multi-modal in-context demonstration selection, leveraging unlabeled data to improve few-shot ICL for MLLMs. Experiments on eight benchmarks show consistent gains in label-scarce regimes.

This paper introduces a geometry-aware adversarial attack framework that targets relational structure in contrastive embedding manifolds, showing that verification systems like Markmatch can be severely degraded by distorting pairwise similarities rather than decision boundaries.

This paper investigates converting pretrained GPT-2 into a time-reasoning model using η-pseudo-unitary operator dynamics, providing mathematical foundations and key findings on PT-breaking transitions and reversible/irreversible sequences.

This paper investigates whether language models encode a structured, human-interpretable consciousness spectrum in their embedding spaces, showing that sentences form a navigable manifold from lower to higher states, with implications for model guidance and alignment.