We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Reducing Retraining by Recycling Parameter-Efficient Prompts

Abstract: Parameter-efficient methods are able to use a single frozen pre-trained large language model (LLM) to perform many tasks by learning task-specific soft prompts that modulate model behavior when concatenated to the input text. However, these learned prompts are tightly coupled to a given frozen model -- if the model is updated, corresponding new prompts need to be obtained. In this work, we propose and investigate several approaches to "Prompt Recycling'" where a prompt trained on a source model is transformed to work with the new target model. Our methods do not rely on supervised pairs of prompts, task-specific data, or training updates with the target model, which would be just as costly as re-tuning prompts with the target model from scratch. We show that recycling between models is possible (our best settings are able to successfully recycle $88.9\%$ of prompts, producing a prompt that out-performs baselines), but significant performance headroom remains, requiring improved recycling techniques.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2208.05577 [cs.CL]
  (or arXiv:2208.05577v1 [cs.CL] for this version)

Submission history

From: Brian Lester [view email]
[v1] Wed, 10 Aug 2022 22:10:53 GMT (276kb,D)

Link back to: arXiv, form interface, contact.