Do we still need Human Annotators? Prompting Large Language Models for Aspect Sentiment Quad Prediction

Benchmark Model Rank Results
aspect-based-sentiment-analysis-absa-on-asqpGemma-3-27B (50-shot, self-consistency learning)#7F1 (R15): 41.74F1 (R16): 51.54
aspect-based-sentiment-analysis-absa-on-asqpGemma-3-27B (10-shot, self-consistency learning)#8F1 (R15): 39.95F1 (R16): 46.23
aspect-based-sentiment-analysis-absa-on-tasdGemma-3-27B (50-shot, self-consistency learning)#5F1 (R15): 62.12F1 (R16): 68.53
aspect-based-sentiment-analysis-absa-on-tasdGemma-3-27B (10-shot, self-consistency learning)#8F1 (R15): 54.37F1 (R16): 66.75