Search papers, labs, and topics across Lattice.
The University of Hong Kong
2
0
3
0
Sparse autoencoders reveal that a Transformer agent's decision-making strategies can be interpreted without explicit rule classification, uncovering hidden concepts that drive behavior.
Forget brittle sim-to-real transfer; CoRMA learns a compact, reusable 6D "contact context" that lets robots adapt to force-dominant assembly tasks on the fly.
