Search papers, labs, and topics across Lattice.
2
0
4
96
Ditch the task-specific verifier: energy-based fine-tuning (EBFT) lets you directly optimize sequence-level behavior in LMs, beating SFT and matching RLVR in downstream tasks.
Language model capabilities are surprisingly stable over time for most tasks, except for math reasoning, which continues to advance, offering a way to reliably translate compute budgets into performance expectations.
