Skip to content

llm-retriever-base

text-embedding

intfloat · intfloat · released

About

LLM-Retriever-Base is intfloat's compact (0.11B) open-weight dense retriever for embedding queries and passages into a shared space — designed for retrieving in-context examples and documents to feed LLMs.

Architecture

Type
encoder
Parameters
109M

LLM-Retriever base: a BERT-base-scale (0.11B) dense retriever that embeds queries and passages into a shared vector space for in-context example / document retrieval. Open-weight, transformers library.

Memory

Weights (BF16)
0.22 GB
Activation estimate
0.10 GB

Pricing

Free — open weights

Self-host on your own GPU. The calculator surfaces GPU-hours cost on the hardware page instead of an API price.

Provenance

License
mit
Last verified
2026-06-22
embeddingretrievaldense-retrieveropen-weight