splade bert tiny nq

BERT

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released August 20191K context≈ 384 words

A compact, sparse retrieval specialist that punches above its weight for keyword-focused search tasks. Built on BERT's tiny architecture, it trades raw capacity for speed and efficiency, producing sparse embedding vectors that highlight the most relevant terms rather than dense representations. Its 512-token context ceiling and minimal parameter count make it fast but limit its ability to handle long documents or nuanced semantic reasoning.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Long Context

splade bert tiny nq

BERT

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released August 20191K context≈ 384 words

A compact, sparse retrieval specialist that punches above its weight for keyword-focused search tasks. Built on BERT's tiny architecture, it trades raw capacity for speed and efficiency, producing sparse embedding vectors that highlight the most relevant terms rather than dense representations. Its 512-token context ceiling and minimal parameter count make it fast but limit its ability to handle long documents or nuanced semantic reasoning.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Long Context

Glossary

ArchitectureThe underlying structural design of a neural network that defines how data flows through layers and components.BERTA foundational neural network architecture designed to understand the meaning of words in context by learning from large amounts of text.EmbeddingA dense numerical vector that represents a word, sentence, or concept in a high-dimensional space.Parameter CountThe total number of adjustable weights in a model; more parameters generally mean more capacity to learn, but also require more computing power.ReasoningThe model's ability to work through multi-step logical problems and provide justified answers rather than just pattern-matching.RetrievalThe process of finding and returning relevant documents or information from a database based on a query.Semantic ReasoningHigh-level planning that operates at a coarser timescale to propose actions and symbolic arguments.Sparse RetrievalA search method that represents text as a high-dimensional vector with mostly zeros, focusing on keyword matching and exact term overlap.TokenA small unit of text (a word, subword, or punctuation mark) that a language model breaks input into for processing.

Capabilities

Capabilities

Use Case Fit

Glossary