Ground Truth
Cross-source consensus on Ground Truth from 1 sources and 5 claims.
1 sources · 5 claims
How it works
Preparation
Risks & contraindications
Evidence quality
Highlighted claims
- Ground truth labels were created through human-in-the-loop regex bootstrapping. — Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction
- Canonicalization mapped heterogeneous grade descriptions to standardized labels. — Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction
- The regex process prioritized explicit diagnostic headers over summary or clinical history mentions. — Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction
- Reproducibility is limited because the reports contain protected health information and are not publicly available. — Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction
- Most of the cohort was manually audited, and the resulting dataset was treated as a high-fidelity ground truth composite. — Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction