What Is ChatGPT?
ChatGPT is a large language model chatbot developed by OpenAI based on GPT=3.5 (Click to learn more). It has a remarkable ability to interact in a conversational dialogue form and provide responses that can appear surprisingly human. Large language models perform the task of predicting the next word in a series of words.
Is ChatGPT Free To Use?
The use of ChatGPT is currently free during the " research preview" time. The chatbot is currently open for users to try out and provide feedback on the responses so that the AI can become better at answering questions and to learn from its mistakes. However, ChatGPT is envisioned as a tool that the public will eventually have to pay to use.
How Can ChatGPT Be Used?
ChatGPT can write code, poems, songs, and even short stories in the style of a specific author. The expertise in following directions elevates ChatGPT from an information source to a tool that can be asked to accomplish a task. This makes it useful for writing an essay on virtually any topic. ChatGPT can function as a tool for generating outlines for articles or even entire novels. It will provide a response for virtually any task that can be answered with written text.
Limitations
●ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as: (1) during RL training, there' s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
●ChatGPT is sensitive to tweaks (微调) to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase (重新措辞), it can answer correctly.
●While we' ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We' re using the Moderation API to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now. We' re eager to collect user feedback to aid our ongoing work to improve this system.
