Systematic assessment of text summarization methods for biomedical literature from frequency methods to language models.

Ce dépôt n'a pas de licence : ses auteurs gardent tous leurs droits. Lisez-le à la source.

• manuscript/build-docx.sh — Shell, 27 lignes
• scripts/add_publication_dates.py — Python, 216 lignes
• scripts/check_data_leakage_completion.py — Python, 1 496 lignes
• scripts/check_data_leakage_cutoffs.py — Python, 584 lignes
• scripts/compute_z_norm_unweigthed_score.R — R, 134 lignes
• scripts/download_models.py — Python, 316 lignes
• scripts/fetch_training_cutoffs.py — Python, 1 321 lignes
• scripts/generate_games_howell_pvalues.py — Python, 446 lignes
• scripts/generate_heatmap_data.py — Python, 265 lignes
• scripts/generate_metric_correlations.py — Python, 122 lignes
• scripts/generate_offline_evaluation.py — Python, 784 lignes
• scripts/generate_plots_manuscript.R — R, 1 039 lignes
• scripts/human_evaluation_analysis.py — Python, 313 lignes
• scripts/human_evaluation_server.py — Python, 1 216 lignes
• scripts/human_evaluation_violin.R — R, 170 lignes
• scripts/human_vs_automatic_correlation.py — Python, 984 lignes
• scripts/length_compliance.R — R, 164 lignes
• scripts/llm_judge_eval.py — Python, 341 lignes
• src/data_models/__init__.py — Python, 3 lignes
• src/data_models/evaluation.py — Python, 101 lignes
• src/data_models/paper.py — Python, 23 lignes
• src/data_models/status.py — Python, 9 lignes
• src/llm_apis/__init__.py — Python, 18 lignes
• src/llm_apis/anthropic_client.py — Python, 147 lignes
• src/llm_apis/base_client.py — Python, 274 lignes
• src/llm_apis/config.py — Python, 90 lignes
• src/llm_apis/exceptions.py — Python, 11 lignes
• src/llm_apis/huggingface_client.py — Python, 256 lignes
• src/llm_apis/local_client.py — Python, 183 lignes
• src/llm_apis/mistral_client.py — Python, 207 lignes
• src/llm_apis/ollama_client.py — Python, 149 lignes
• src/llm_apis/openai_client.py — Python, 203 lignes
• src/llm_summarization_benchmark/__init__.py — Python, 1 ligne
• src/llm_summarization_benchmark/benchmark.py — Python, 1 249 lignes
• src/llm_summarization_benchmark/check.py — Python, 24 lignes
• src/llm_summarization_benchmark/goldstandard_proxy_validation.py — Python, 319 lignes
• src/llm_summarization_benchmark/metrics.py — Python, 794 lignes
• src/llm_summarization_benchmark/visualization.py — Python, 1 851 lignes
• src/utilities/__init__.py — Python, 6 lignes
• src/utilities/environment.py — Python, 79 lignes
• src/utilities/logger.py — Python, 96 lignes
• src/utilities/summarization.py — Python, 256 lignes
• README.md — Texte, 279 lignes