Search papers, labs, and topics across Lattice.
5
1
8
25
This work defines the scaling inflection point as the per-session budget where marginal Elo gains match the independent-sampling reference, and proposes Elo-per-token analysis, which tracks the best solution found at each token budget and uses a Bradley-Terry model to aggregate within-task orderings into Elo ratings across tasks with different score scales.
Switch Distillation not only enhances reasoning performance by up to 71% but also preserves factual recall, challenging the conventional trade-off in knowledge distillation.
Discarding irrelevant reasoning tokens can make language models up to 3x faster at test time without sacrificing performance.
By making environment design a learnable process, SPADE unlocks a new frontier in self-improvement for language agents, leading to substantial performance gains across diverse tasks.
Instant AI assistants are now feasible on smartwatches: 8M-parameter models can kickstart responses locally, hiding cloud latency with surprisingly high quality.
