Describe what the 'query:' and 'passage:' prefixes do for models like E5 and BGE

Flashcard·Easy·4.0 · 0·~30s

Attempt it

Your answer

Many open-weight embedding models like **multilingual-e5-large** and **BGE-M3** require an instruction prefix on every input. Queries get 'query: ' prepended; documents (passages) get 'passage: ' prepended. The prefixes look like ordinary preprocessing strings, but they are not. They are part of the **training contract**. During contrastive training, the model saw queries with the 'query: ' prefix and passages with the 'passage: ' prefix in every batch. The model learned to project the two sides into slightly different sub-regions of the embedding space, capturing the asymmetry between short intent-bearing queries and longer content-bearing passages. At inference time you have to follow the same contract. Embed queries with 'query: ' and passages with 'passage: '. Mix them up or skip the prefix on one side and the geometry shifts: the model embeds the text as if it were generic. Which drops retrieval recall by 5 to 15 percent compared to using the prefixes correctly. This is one of the most common silent bugs in production embedding pipelines. There is no error, no warning, no API rejection. The vectors are still valid floats, cosine similarities still produce numbers, but the recall benchmark would show the gap if you measured it.

**Instruction prefixes** are part of the model's training contract, not optional preprocessing. Models like multilingual-e5-large, multilingual-e5-base, BGE-large-en-v1.5, BGE-M3, and several instruction-tuned embedders were contrastive-trained with explicit asymmetric prefixes on queries vs passages. The purpose is to encode the **asymmetry** between the two sides of retrieval. Queries are typically short, intent-bearing, question-shaped; passages are longer, content-bearing, declarative. A model that projects both into a single uniform geometric space wastes capacity on the asymmetry. A model that uses prefixes to signal which side the input represents can project queries and passages into matched but distinct sub-regions of the same space. This with the contrastive loss aligning matched pairs across the sub-regions. The failure mode is silent. If you skip the prefix on one side but apply it on the other, you embed the un-prefixed input as if it were generic content. The vectors are still valid floats and cosine similarities still produce numbers in the expected range. The geometry has shifted by a few degrees, enough to drop retrieval recall by 5 to 15 percent depending on the model and benchmark. No error fires, no log line appears, nothing fails loudly. The only way to detect the bug is a recall benchmark on a labeled eval set. This is the single most common silent bug in 2026 production RAG pipelines. Teams onboard a new embedding model, copy the API call shape from a tutorial, forget to add the prefix on the query side. Absorb a percentage points of recall hit they never measure.

Quick reference

TL;DR

The prefixes are part of the training contract, they tell the encoder to project queries and passages into matched but distinct sub-regions of the embedding space; forgetting them silently drops recall.

Memory aid

Easy to grasp

Imagine a sommelier who studied under two strict instructors. One taught them how to taste short, sharp questions like a customer at the bar. The other taught them how to taste long, descriptive passages like a wine label. The sommelier was always told before each tasting which mode to switch into. Now picture you walk up and hand them a sip without saying which mode. They guess, and they often guess wrong, and the pairings they recommend at the end of the night no longer line up. The little phrase like query colon or passage colon is that whispered hint. Skip it and the recommendations quietly fall apart, with no error message, no warning, only worse matches.

Key concepts

Embeddings

Detailed answer & concept explanation

~6 min read

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example. Click to expand.

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example.

Quick scan · 2–4 min

Everything important, quickly.

5 to 7 min: training-contract framing + query-passage asymmetry + silent-recall failure mode + detection via recall benchmark + 2026 model landscape.

Real-world examples

Real products, models, and research that use this idea.

multilingual-e5-large from Microsoft requires 'query: ' and 'passage: ' prefixes for asymmetric retrieval.
BGE-large-en-v1.5 from BAAI requires task-specific instructions like 'Represent this sentence for searching relevant passages: ' on the query side.
BGE-M3 supports multiple retrieval modes (dense, sparse, multi-vector) and each has its own prefix convention.
e5-mistral-7b-instruct uses task-specific instructions like 'Instruct: Given a question, retrieve relevant passages. Query: ...'.
OpenAI text-embedding-3-large and Voyage v3-large abstract any prefix behind the API, you do not configure it yourself.

Interviewer follow-ups

What an interviewer would ask next. Try answering before peeking at the approach.

QHow do you verify the right prefix to use for a given model?

QWhat is the geometric intuition behind why misapplied prefixes drop recall by 5 to 15 percent?

QHow would you migrate from a no-prefix to a with-prefix embedding model in production?

Don't say this

Red flags and common mistakes that signal junior thinking. Click to expand.

Most common mistake

Treating the prefix as optional preprocessing. It is part of the training contract, omitting it produces silent recall degradation with no error.

60-second cheat sheet

The night-before-the-interview bullets. Scan these on the way to the call.

References

Primary sources. Skim if you want the original framing.