Get Started
Home
Topics
Search
Library
Research questionHow can climate disclosure classifiers remain reliable when documents shift across sources?Annual reports, press releases, and earnings calls differ in length, purpose, and writing style. Adaptation that performs well on one source may therefore lose effectiveness when applied to another.
AI
Business
Earnings Call
Evaluation & Benchmarks
Finance
Machine Learning
Natural Language Processing
Latest papersRecent research connected to this question, newest first.What Transfers Under Source Shift? Definitions, Examples, and Fine-Tuning for Climate Disclosure ClassificationThe study examines definitions, few-shot examples, and fine-tuning across eleven open- and closed-source LLMs using two corpora with the same label space but different document sources. It reports that definitions transfer most consistently when their granularity matches the target text, while retrieval-based examples and LoRA fine-tuning can lose their in-source advantage under source shift; randomly selected examples are more robust than their in-source results suggest.research paper · Sep 2, 2026
Related questions
How can authorship attribution remain reliable for long-form LLM text across languages and distribution shifts?How can AI-generated text detectors remain reliable under distribution shift?How can pediatric chest X-ray classifiers preserve usable decisions across countries when calibration and source thresholds shift?How can classification models adapt to removed labels without retraining on the original data?