Text embeddingA list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.Full definition → are numerical representations of text that systems use for search, grouping and ranking. A model that represents text well for similarity comparisons may perform differently when asked to group documents or reorder search results. The original MTEB study addressed that selection problem by evaluating 33 text-embedding models across eight task types, rather than judging them on a single kind of test.
The abstract published by Cohere at Cohere's research site reports that no method dominated every task. Its useful finding is that embedding quality depends on the intended job: a strong similarity score is insufficient evidence that a model is the best choice for a different application.
Terms in this piece · Glossary
embedding — A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
postCohere releases open-weights translation model North Small Translate
articleThe Data Provenance Initiative A Large Scale Audit Of Dataset Licensing And Attribution In Ai 2023 10 25
