API Generating Language Model: Difference between revisions
Jump to navigation
Jump to search
m (Text replacement - "** ..." to "** …") |
m (Text replacement - ".↵----" to ". ----") |
||
| Line 17: | Line 17: | ||
** … | ** … | ||
* <B>See:</B> [[API Call]], [[LLM for Programming]], [[Code Generation]]. | * <B>See:</B> [[API Call]], [[LLM for Programming]], [[Code Generation]]. | ||
---- | ---- | ||
---- | ---- | ||
[[Category:Concept]] | [[Category:Concept]] | ||
Revision as of 23:03, 26 November 2023
A API Generating Language Model is a software-generating language model specialized for generating API calls from natural language instructions.
- Context:
- It can be trained on API documentation and associated natural language descriptions.
- It takes a specification and outputs an API call with the right components filled in.
- It focuses on producing valid API calls over full programs.
- It is evaluated using AST matching against reference APIs.
- It is optimized for invoking APIs as tools.
- …
- Example(s):
- Gorilla LLM ().
- Anthropic's Claude can produce API calls for a variety of provided APIs.
- Toolformer LLM.
- …
- Counter-Example(s):
- A Chatbot LLM is not optimized for API generation.
- An LLM for Mathematical Reasoning focuses on math not coding.
- …
- See: API Call, LLM for Programming, Code Generation.