Search papers, labs, and topics across Lattice.
3
0
6
17
Filtering out fit sequences during fine-tuning can boost post-training performance by up to 17%, reshaping how we approach model training for RL applications.
Recurrent Transformers let you trade model depth for width, slashing KV cache memory footprint and inference latency without sacrificing performance.
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.
