Difficulty-Based Preference Data Selection Improves DPO and RLHF Sample Efficiency on SQuTR

Abstract

Abstract: This report synthesises findings from 13 peer-reviewed papers addressing the following research question: What is the impact of dataset size on the sample efficiency of DPO versus RLHF methods when aligning LLMs on SQuTR with varying input noise distributions. Aligning large language models (LLMs) with human preferences is a critical challenge in AI research. While methods like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) are widely used, they often rely on large, costly preference. 14 claims were extracted from source literature; 1 was independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 4.8/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research Question

What is the impact of dataset size on the sample efficiency of DPO versus RLHF methods when aligning LLMs on SQuTR with varying input noise distributions?

Verification Level

Descriptive public verification status only; aggregate claim counts are public, but individual claim records are not exposed here.

Quality Tier

Descriptive public triage only; this tier does not alter current publication or DOI behavior.

Quality Dimensions

Automated triage signals derived from public fields; not human peer review or independent validation.

Correction Record

Public corrections are additive records. Current status does not claim the synthesis is error-free.

Provenance