ResearchBenchmarksSummarization on Human EvaluationFollow4.4CoherenceGloSA-sum3.9844.0924.24.308Feb 10, 2026Evaluation ResultsMethodMethodLinksCoherenceInformativenessConcisenessAverage ScoreGloSA-sum2026.024.44.34.24.3GPT-4mode=Prompt-Summode=Prompt-Sum2026.024.34.44.24.3Claude-3mode=Summode=Sum2026.024.24.34.14.2LLM Summode=RAGmode=RAG2026.024.14.24.14.13LLaMAtraining_strategy=Fine...training_strategy=Fine-tuned2026.0244.144.03