Google's RPTune Paper: Catalog Curation + Post-Training Boosts LLM Search Accuracy by 31.4 Points
_reachsumit · x · 2026-10-02
Google researchers (Chuxuan Hu et al.) introduce RPTune, a framework for LLM catalog search aimed at small merchants whose catalogs fit inside a long context window.
The key insight: stuffing the full catalog into context doesn't mean the model uses it well, since LLMs exploit long contexts unevenly. RPTune attacks this from two complementary angles:
- Learned catalog curation: an encoder-reorganizer curator orders and prunes products guided by downstream LLM feedback;
- LLM post-training: using automatically generated, catalog-grounded supervision with a context-relative reward, with curated catalogs in turn improving post-training effectiveness.
Evaluated on 7 real merchants across retail verticals with 100 complex conversational queries each, curation yields gains of up to 31.4 percentage points and post-training adds 10.3 points on average, consistent across proprietary and open-weight LLMs.
Related event: Google's RPTune boosts LLM catalog search accuracy by up to 31.4 points(2 posts)→
More from Research
- New video traces how adversarial objectives evolved beyond GANs and self-play — manicman1999 · 2026-10-02
- Peer-reviewed paper questions whether gene expression model benchmarks are informative — simocristea · 2026-10-02
- Oxford VGG Unveils SynCity 3000, Generating Globally Coherent Scene-Scale 3D Worlds — rsasaki0109 · 2026-10-02
- Christian Szegedy revisits 2019 interview: his 'crazy' timelines for AI math and code — ChrSzegedy · 2026-10-02
- Omni-Embed-Mini: A 0.9B Embedder Adds Five Modalities Without Touching Text Weights — _reachsumit · 2026-10-02
- Apple paper: structured selection-based reasoning cuts search agent latency by 90% — _reachsumit · 2026-10-02