Manifold-based Verbalizer Space Re-embedding for Tuning-free Prompt-based Classification
This work addresses the need for efficient, tuning-free classification methods in prompt-based learning, particularly for hyper-scale language models, though it is incremental in improving existing verbalizer embedding techniques.
The paper tackles the problem of tuning-free prompt-based classification by proposing a manifold-based space re-embedding method (LLE-INC) for verbalizer embeddings, which achieves performance on par with tuned automated verbalizers and enhances prompt-based tuning by up to 3.2% with parameter updating.
Prompt-based classification adapts tasks to a cloze question format utilizing the [MASK] token and the filled tokens are then mapped to labels through pre-defined verbalizers. Recent studies have explored the use of verbalizer embeddings to reduce labor in this process. However, all existing studies require a tuning process for either the pre-trained models or additional trainable embeddings. Meanwhile, the distance between high-dimensional verbalizer embeddings should not be measured by Euclidean distance due to the potential for non-linear manifolds in the representation space. In this study, we propose a tuning-free manifold-based space re-embedding method called Locally Linear Embedding with Intra-class Neighborhood Constraint (LLE-INC) for verbalizer embeddings, which preserves local properties within the same class as guidance for classification. Experimental results indicate that even without tuning any parameters, our LLE-INC is on par with automated verbalizers with parameter tuning. And with the parameter updating, our approach further enhances prompt-based tuning by up to 3.2%. Furthermore, experiments with the LLaMA-7B&13B indicate that LLE-INC is an efficient tuning-free classification approach for the hyper-scale language models.