With more than 100 million users worldwide, ChatGPT is undoubtedly one of the most popular AI-powered chatbots of all time. In a recent blog post, OpenAI announced that ChatGPT is getting a feature that will let users share their conversations with the chatbot.
Dubbed ‘Shared Links’, the new feature eliminates the need to copy-paste or screenshot replies by generating a unique URL for every conversation you have with ChatGPT. Any messages you send after creating the link won’t be shared with others, so users who want to update shared content will have to delete and reshare the link.
OpenAI says the feature is available to ChatGPT Plus subscribers and will be rolled out to normal users in the coming days, so you might have to wait for some time before it appears on your account. Also, the feature is currently limited to the web version with support for the iOS app still in development
Generating links to ChatGPT conversations is really easy. To do so, simply hover over the conversation name that appears on the left side of the screen and click on the share button that appears. A new window will now appear with the conversation you want to share.
On the bottom, you will find the ‘Copy Link’ button that copies the generated URL to your clipboard. Users can also name the conversation before sharing and add their name by pressing the three-dot button on the right side. In case you want to delete the URL to the shared conversation, just press the three-dot menu after copying the link can you will find an option that lets you do so.
Keep in mind that anyone with access to the shared link can view and continue the linked conversation, so make sure you don’t share any sensitive content in shared conversations.
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 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, can answer correctly.
The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.1,2
Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
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.
Better
You must be logged in to post a comment.