Pair2Score: Pairwise-to-Absolute Transfer for LLM-Based Essay Scoring

TL;DR

Pair2Score introduces a two-stage framework that effectively transfers pairwise comparison data into absolute essay scores using LLaMA, improving scoring accuracy across multiple traits.

Contribution

It presents a novel transfer learning approach that leverages pairwise comparisons to enhance absolute scoring models with parameter-efficient LLaMA adaptation.

Findings

Transfer variants improve QWK over baseline for all traits.

One-epoch pairwise training transfers more reliably than extended training.

Transfer configuration choice critically impacts scoring performance.

Abstract

Many scoring applications require absolute predictions, while pairwise comparisons can provide a simpler learning objective. We present Pair2Score, a two-stage learning framework that transfers pairwise comparisons into absolute scoring with parameter-efficient LLaMA adaptation. Stage 1 trains a directional Siamese ranker on pairwise comparisons derived from absolute trait labels; Stage 2 trains an absolute predictor using configurable transfer strategies (warm-start and embedding-fusion variants). We evaluate on rubric-aligned Automated Essay Scoring (AES) traits (grammar, vocabulary, syntax) under a five-fold protocol that co-rotates held-out fold and random seed. At the trait level, the best-performing transfer variant improves quadratic weighted kappa (QWK) over an absolute-only baseline for all three traits. However, not all transfer configurations help: a one-epoch pairwise stage transfers more reliably than extended pairwise training, and transfer configuration -- not just the inclusion of a pairwise stage -- determines whether downstream scoring benefits.