Pith. sign in

Frustratingly simple retrieval improves challenging, reasoning-intensive benchmarks

2 Pith papers cite this work. Polarity classification is still indexing.

2 Pith papers citing it

fields

cs.CL 1 cs.IR 1

years

2026 2

representative citing papers

Procedural Knowledge at Scale Improves Reasoning

cs.CL · 2026-04-01 · conditional · novelty 6.0

Retrieving compact procedural hints from 32M subquestion–subroutine pairs improves reasoning-model accuracy on math, science, and coding benchmarks beyond compute-matched test-time scaling.

citing papers explorer

Showing 2 of 2 citing papers.

  • PIXELRAG: Web Screenshots Beat Text for Retrieval-Augmented Generation cs.IR · 2026-06-01 · unverdicted · none · ref 6

    PixelRAG shows that operating RAG entirely over web screenshots outperforms text-based retrieval on NQ, SimpleQA, MMSearch, LiveVQA, and MoNaCo, with up to 18.1% accuracy gains and 3x token savings via image compression.

  • Procedural Knowledge at Scale Improves Reasoning cs.CL · 2026-04-01 · conditional · none · ref 20

    Retrieving compact procedural hints from 32M subquestion–subroutine pairs improves reasoning-model accuracy on math, science, and coding benchmarks beyond compute-matched test-time scaling.