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

Download:

Current browse context:

cs.CL

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 > Machine Learning

Title: Prompt Injection: Parameterization of Fixed Inputs

Abstract: Recent works have shown that attaching prompts to the input is effective at conditioning Language Models (LM) to perform specific tasks. However, prompts are always included in the input text during inference, thus incurring substantial computational and memory overhead. Also, there is currently no straightforward method of utilizing prompts that are longer than the maximum input length of the LMs without incurring additional costs during inference. We propose Prompt Injection (PI), a novel formulation of injecting the prompt into the parameters of an LM to be an efficient alternative to attaching fixed prompts to the input. We show that in scenarios with long fixed prompts, PI can be up to 280 times more efficient in terms of total FLOPs than previous approaches. We further explore methodologies for PI and show promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions. Through these explorations, we show that PI can be a promising direction for conditioning language models, especially in scenarios with long and fixed prompts.
Comments: PING results in Table 2 updated (bug fixed)
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as: arXiv:2206.11349 [cs.LG]
  (or arXiv:2206.11349v2 [cs.LG] for this version)

Submission history

From: Eunbi Choi [view email]
[v1] Tue, 31 May 2022 08:43:07 GMT (2133kb,D)
[v2] Fri, 15 Jul 2022 07:15:31 GMT (2133kb,D)

Link back to: arXiv, form interface, contact.