50Off-the-shelf embeddings retrieve poorly on your domain. How do you improve retrieval accuracy?▼hardCohereGleanHarvey2 replies◆ premiumA model that tops MTEB can still flop on your jargon-heavy corpus, and most teams jump straight to the expensive fix. There is a ladder, and the cheap rungs are the ones people skip. Here is the order to climb it.Open full answer →
73When do you fine-tune a reranker on your own data, and how do you build the training set?▼hardCohereGleanMicrosoft1 replies◆ premiumAn off-the-shelf cross-encoder is general; your domain has jargon and relevance rules it never saw. The signal is knowing when fine-tuning pays off, how to mine hard negatives, and how to avoid training a reranker that just memorizes your retriever's mistakes.Open full answer →
91You're training embeddings with contrastive/triplet loss. How do you choose pairs, the margin, and negatives?▼hardGoogleMetaCohere2 replies◆ premiumMetric learning lives or dies on the pairs you feed it. Random negatives teach almost nothing, and the margin and mining strategy decide whether the embeddings are any good. Here is how the choices interact.Open full answer →
