Rohan Thomas

Cross-Model Stitching Transfers Answers, Not Computation

Activation stitching describes the technique of grafting one model’s hidden state into another’s through a learned map. Its reasoning is grounded in universality, under which shared circuitry between models is assumed to enable seamless transfer. Prior work validates this techniq...

Cross-Model Stitching Transfers Answers, Not Computation

Activation stitching describes the technique of grafting one model’s hidden state into another’s through a learned map. Its reasoning is grounded in universality, under which shared circuitry between models is assumed to enable seamless transfer. Prior work validates this techniq...