This repository was archived by the owner on Jan 2, 2025. It is now read-only.
Search with hypothetical documents #711
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ggordonhall
force-pushed
the
gabriel/blo-1183-diverse-search-queries
branch
from
262a688 to
0a5cf8e
Compare
rsdy
approved these changes
Jul 4, 2023
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Our semantic search pipeline sometimes fails to return relevant information. The queries that the agent generates do not overlap at all with the correct documents. For example, for the user query
which ml library do we use?bloop might make a semantic search forml libraryormachine learning library, neither of which overlap withortoronnxruntime.This PR implements a popular approach to tackling this recall problem. As well as making a semantic search with the query, use an LLM to generate a set of hypothetical documents that could answer the query and search w.r.t. those too. https://arxiv.org/abs/2212.10496
We take the user query and ask GPT-3.5 to generate three code snippets in a variety of languages that could answer it. We then batch search Qdrant for all of those queries at once (very open to more elegant implementations of this).