Search papers, labs, and topics across Lattice.
4
0
2
5
Unbiased sampling of source prefixes in TLMs can reduce runtime by several orders of magnitude while maintaining accuracy in estimating target prefix probabilities.
Surprisal theory's reliance on arbitrary tokenization schemes undermines its validity, but this framework offers a way to fix it.
Repurpose existing language models for entirely new output formats like bytes, words, or even DNA sequences, all without retraining, by wrapping them in a finite-state transducer.
Forget simple probability averaging: a byte-level sequential Monte Carlo algorithm unlocks better language model ensembles by enabling sophisticated aggregation strategies and handling mismatched vocabularies.
