# GPT

## Precursor

1. [Proximal Policy Optimization](https://openai.com/research/openai-baselines-ppo) (PPO) - an RL algorithm, PPO is better than state-of-the-art approaches while being much simpler to implement and tune and is the default reinforcement learning algorithm at OpenAI.
2. [Learning from human preference](https://openai.com/research/learning-from-human-preferences) (human in the loop) - a method used to infer what humans want by being told which of two proposed behaviors is better.
3. [instructGPT](https://openai.com/research/instruction-following) - arguably better at following user intentions than GPT-3 while also making them more truthful and less toxic, using human in the loop.

## Articles

1. [what is chatGPT doing and why does it work?](https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/) explaining next word prediction in detail.
2. [Karpathy on building GPT](https://www.youtube.com/watch?v=kCc8FmEb1nY\&t=191s)
3. [Is DPO Superior to PPO for LLM Alignment](https://arxiv.org/pdf/2404.10719)? A Comprehensive Study - "PPO is able to surpass other alignment methods in all cases and achieve state-of-the-art results in challenging code competitions."

## Competitions

1. GPT 4 [Hackathon code results](https://docs.google.com/spreadsheets/d/1tmfn8jKb7T1x7PpyO7rD023tH2zc_WDg_OHh0aVXIrw/edit#gid=174517450)
2. [LangChain Gen Hackathon](https://docs.google.com/spreadsheets/d/1GqwPo1FpAbe_awmNZW5ZMH69yc5QtEr7ZYw-ckaz_mQ/edit#gid=795016726)

## Tools

1. Sentence Embeddings
   1. [sentence embedding for semantic search](https://github.com/Muennighoff/sgpt)
   2. [GPT 3 Dense sentence embeddings](https://medium.com/@nils_reimers/openai-gpt-3-text-embeddings-really-a-new-state-of-the-art-in-dense-text-embeddings-6571fe3ec9d9)

## Virtual assistants

1. [flowGPT](https://flowgpt.com/) - has many bots, prompts.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://www.mlcompendium.com/generative-ai/gpt.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.