A Systematic Review on Automatic Prompt Learning

Prompts direct the behavior of a model by conditioning its outputs on carefully designed instructions and examples, similar to setting the trajectory of an arrow before release. In a broader sense, prompt learning is the research area that aims to solve downstream tasks by directly leveraging the knowledge acquired by language models at pre-training time, removing the necessity for costly fine-tuning stages with potentially different objective functions. While manual prompt engineering has enabled both small and large language models to achieve superhuman performance on various benchmarks, it remains a demanding and less-than-ideal method. Recently, the field has shifted towards automating the search for prompts that effectively elicit the desired model responses. In this thesis we presents a systematic review of prompt learning for pre-trained language models operating on textual inputs, with a particular focus on automatic methods. We critically analyze existing publications and organize them into a novel taxonomy, highlighting essential elements for practical application. We also performed practical experiments and performed performance tests applying the best automatic prompt learning techniques. Finally, we discussed promising directions for future research. Our curated repository of annotated papers, continuously updated, is available at https://github.com/disi-unibo-nlp/awesome-prompt-learning.