arXiv:2604.11582v3 Announce Type: replace-cross
Abstract: Standard subword tokenization methods fragment numbers inconsistently, causing large language models (LLMs) to lose positional and decimal structure – a primary driver of errors in arithmetic a…
arXiv:2604.11582v2 Announce Type: replace-cross
Abstract: Standard subword tokenization methods fragment numbers inconsistently, causing large language models (LLMs) to lose positional and decimal structure – a primary driver of errors in arithmetic a…
arXiv:2604.11582v1 Announce Type: new
Abstract: Standard subword tokenization methods fragment numbers inconsistently, causing large language models (LLMs) to lose positional and decimal structure – a primary driver of errors in arithmetic and scienti…