aiqbench takeAI Products

Perplexity's new pplx-embed-v2-context-9b-preview introduces a significant shift in RAG training by using full-document context to verify answers rather than relying on isolated gold passages. This approach addresses limitations in binary labeling and chunking dependency, leveraging a query-aware teacher model to distill retrieval signals without adding inference latency. The release offers a deployable preview for researchers aiming to improve the robustness of retrieval-augmented generation pipelines.

Source coveringMarkTechPost

What this isA model's reading of the article and related coverage, written for aiqbench. Not the publisher's words.

Perplexity Releases pplx-embed-v2-context-9b-preview: A Contextual Embedding Model That Retrieves Answers and Their Supporting Evidence

MarkTechPostOctober 1, 20265 min read

Perplexity Research and turbopuffer have released pplx-embed-v2-context-9b-preview, a contextual embedding model for RAG pipelines. Each chunk is embedded with the full document in view. The real change is the training signal. The model learns to retrieve the answer along with the context needed to verify it, not one ‘gold passage.’ P Is it deployable? Yes, […] The post Perplexity Releases pplx-embed-v2-context-9b-preview: A Contextual Embedding Model That Retrieves Answers and Their Supporting Evidence appeared first on MarkTechPost.

Perplexity Research and turbopuffer have released pplx-embed-v2-context-9b-preview, a contextual embedding model for RAG pipelines. Each chunk is embedded with the full document in view. The real change is the training signal. The model learns to retrieve the answer along with the context needed to verify it, not one ‘gold passage.’ P Is it deployable? Yes, as a self-hosted preview. Weights are on Hugging Face under the MIT license. Loading requires transformers>=5.4.0 with trust_remote_code=True. It is not yet on the Perplexity API. The model card notes that weights and interface may change without…

Excerpt shown under the publisher's RSS terms · 92 words · full text and figures are on the publisher's site

Read at MarkTechPost ↗Ask about this article