LIT Prompts av Colarusso

Create, save, and run interconnecting template-based prompts for Large Language Models (LLMs) using direct input along with content from your active browser window.

0 (0 reviews)

28 brukarar

Du treng Firefox for å bruke denne utvidinga

Last ned Firefox og få utvidinga

Last ned fil

Om denne utvidinga

A project of Suffolk University Law School's Legal Innovation & Technology (LIT) Lab, this extension was designed to help students explore the use of Large Language Models (LLMs) and prompt engineering. Users can create and save prompt templates for use by the extension with any OpenAI-compatible API.

Templates can use data from ones active browser window (e.g., selected text or the whole text of a webpage) as well as text input provided by the user. They can also feed into each other, allowing one to create complex interactions. For example: (1) summarize and ask questions of a webpage (e.g., What are the main conclusions of this article? What did so and so say about this or that?); or (2) translate selected text from one language into another. Then draft a reply in the original language.

Vurdert 0 av 0 meldarar

Versjonsnotat for 0.2.0

Bug fixes and enhancements.

- Removed Storage from permissions
- Added Active Tab URL to transcripts
- Add shift+enter as an exception to send on inputs (allows for line breaks in input)
- Added "restart" to Chat workflow
- Show how many characters are in output enhancement
- Added trim() to output to remove leading and trailing blank characters
- Changed copy to clipboard scope to include only most-recent output.
- Encrypt and password protect API Key on export to file
- Disable resize on exports without the scratch pad (single pane outputs
- Improve button appearance for iOS
- Fixed bug with {{scratch}} variable not rendering as string.
- Fixed bug to force repeat of novel questions
- Fixed bug with "Continue Generation" post-run behavior

LIT Prompts av Colarusso

Påkravde løyve:

Valfrie løyve: