Generate an LLM summary
Works best on models with real time data RAG (given the fast pace of developments etc).
Outputs can be surprisingly rich given that this is a relatively short prompt:
Basic prompt template
For this LLM:
{llm-name}
Provide the following:
- Release date
- End of training period
- Strengths
- Weaknesses
- Preceded by
- Performance on benchmarks
- Comparable to
With Markdown Formatting Insturction
For this LLM:
{llm-name}
Retrieve the following:
- Release date
- End of training period
- Strengths
- Weaknesses
- Preceded by
- Performance on benchmarks
- Comparable to
Format the output as a markdown table:
With CSV Formatting Instruction
For this LLM:
{llm-name}
Retrieve the following:
- Release date
- End of training period
- Strengths
- Weaknesses
- Preceded by
- Performance on benchmarks
- Comparable to
Format the output as CSV with this header row:
{header-row-for-consistency}
More Extensive Version With Section Formatting Instructions
For this LLM:
`{llm-name}`
Provide the following and use the following markdown formatting headers as your model to follow in how you structure your output.
# Basic Parameters
## Release date
(When was the LLM released)
## Training period & data
(What is known about the training period it underwent )
## Release Date
(When the model was released)
# Reception
## Benchmark performance
## Critical reception