PACIT: Unlocking the Power of Examples for Better In-Context Instruction Tuning
This work addresses the challenge of making instruction tuning more effective for AI developers and researchers, though it is incremental as it builds on existing in-context instruction tuning methods.
The paper tackles the problem of improving instruction tuning for large language models by proposing PACIT, a method that uses examples to enhance learning, resulting in performance gains of up to 9.16 and 3.14 average ROUGE-L scores on in-domain and out-domain tasks compared to a baseline.
Instruction tuning enhances the instruction following ability of large language models by finetuning with supervised instruction data. Previous work proposes in-context instruction tuning (ICIT) where specific positive or negative examples are incorporated into the prompt for better performance. In this work, we propose PACIT, a simple and effective in-context instruction tuning method, inspired by the pedagogical concept of desirable difficulty. The PACIT method unlocks the power of examples by encouraging the model to actively learn to grasp the distinctions between the positive and negative examples instead of merely reading. The model is expected to first verify the correctness of the provided example according to the task description, which is then set as the condition for generating a better response to the task instance. Our extensive experiments prove the effectiveness of PACIT, outperforming ICIT baseline on both in-domain and out-domain tasks up to 9.16 and 3.14 average ROUGE-L scores, respectively. Moreover, PACIT can notably enhance the performance of instruction tuning even when all positive and negative examples are generated with a self-instruct method.