Cross-Tokenizer LLM Distillation through a Byte-Level Interface
This addresses the challenge of knowledge transfer in language models when tokenizers differ, though it is incremental as it builds on existing distillation methods.
The paper tackled the problem of cross-tokenizer distillation (CTD) by proposing Byte-Level Distillation (BLD), a method that operates at the byte level to transfer knowledge between language models with different tokenizers, achieving competitive or superior performance on benchmarks with models from 1B to 8B parameters.
Cross-tokenizer distillation (CTD), the transfer of knowledge from a teacher to a student language model when the two use different tokenizers, remains a largely unsolved problem. Existing approaches rely on heuristic strategies to align mismatched vocabularies, introducing considerable complexity. In this paper, we propose a simple but effective baseline called Byte-Level Distillation (BLD) which enables CTD by operating at a common interface across tokenizers: the byte level. In more detail, we convert the teacher's output distribution to byte-level probabilities, attach a lightweight byte-level decoder head to the student, and distill through this shared byte-level interface. Despite its simplicity, BLD performs competitively with--and on several benchmarks surpasses--significantly more sophisticated CTD methods, across a range of distillation tasks with models from 1B to 8B parameters. Our results suggest that the byte level is a natural common ground for cross-tokenizer knowledge transfer, while also highlighting that consistent improvements across all tasks and benchmarks remain elusive, underscoring that CTD is still an open problem.