Early-Layer LoRA Fine-Tuning Outperforms Full-Model Tuning for Zero-Shot Cross-Lingual NLI in Niger-Congo Languages

Abstract

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: Does the early-layer LoRA fine-tuning strategy used in Targeted Lexical Injection improve zero-shot cross-lingual natural language inference accuracy on XNLI for Niger-Congo languages outside the. 9 claims were extracted from source literature; 2 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 4.7/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research Question

Does the early-layer LoRA fine-tuning strategy used in Targeted Lexical Injection improve zero-shot cross-lingual natural language inference accuracy on XNLI for Niger-Congo languages outside the Bantu family compared to full-model fine-tuning?

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