ChatGPT could have multiple preset personalities for you to interact with in the future, to help combat its sycophantic personality problem

[Pelican Press 0]

This is the hidden content, please

• Sign In
• or
• Sign Up

ChatGPT could have multiple preset personalities for you to interact with in the future, to help combat its sycophantic personality problem

---

• OpenAI has addressed what went wrong with its sycophantic update
• OpenAI is refining core training techniques to avoid sycophancy in the future
• Its Head of Model Behavior suggests ChatGPT could have multiple personalities going forward

---

OpenAI has addressed the rollback of the most recent update to ChatGPT-4o because of user complaints about its annoying, sycophantic personality. In an

This is the hidden content, please

• Sign In
• or
• Sign Up

OpenAI states:

“We have rolled back last week’s GPT‑4o update in ChatGPT so people are now using an earlier version with more balanced behavior. The update we removed was overly flattering or agreeable – often described as sycophantic.

We are actively testing new fixes to address the issue. We’re revising how we collect and incorporate feedback to heavily weight long-term user satisfaction and we’re introducing more personalization features, giving users greater control over how ChatGPT behaves.”

You may like

The blog post goes into further reasons for why the problem happened, but the crux of the matter is that OpenAI focused too much on short-term feedback, and did not fully account for how users’ interactions with ChatGPT evolve over time.

OpenAI promises to refine its core training techniques and systems prompts to explicitly steer the model away from sycophancy in the future, as well as allow for more user testing before deploying models.

Ask me anything

From reading the blog post, you get the sense that OpenAI feels like it has dropped the ball on this one and is doing all it can to make amends.

As well as issuing the statement, Joanne Jang, its Head of Model Behavior, has just done an

This is the hidden content, please

• Sign In
• or
• Sign Up

, to answer any further questions people may have, specifically about ChatGPT’s personality, sycophancy, and future model behavior. Her replies revealed interesting insights into what the future might hold for ChatGPT’s personality:

Sign up for breaking news, reviews, opinion, top tech deals, and more.

This is the hidden content, please

• Sign In
• or
• Sign Up

user Responsible_Cow2236 asked, “Where do you see the future of model behavior heading? Are we moving toward more customizable personalities, like giving users tools to shape how ChatGPT sounds and interacts with them over time?”

Jang responded by offering the possibility that ChatGPT could come with multiple preset personalities:

“We’ve been working on two things: (1) getting to a default personality that might be palatable for all users to begin with (not feasible but we need to get somewhere) and (2) instead of relying on users to describe / come up with personalities on their own, offering presets that are easier to comprehend (e.g. personality descriptions vs. 30 sliders on traits)”, said Jang.

The idea of preset personalities in ChatGPT is an intriguing one and something I could see combining well with the different voices that ChatGPT advanced voice mode already offers.

As for when we can expect the next update to ChatGPT-4o, OpenAI is not saying anything at the moment beyond “We fell short and are working on getting it right.”

You may also like

This is the hidden content, please

• Sign In
• or
• Sign Up

#ChatGPT #multiple #preset #personalities #interact #future #combat #sycophantic #personality #problem

This is the hidden content, please

• Sign In
• or
• Sign Up

This is the hidden content, please

• Sign In
• or
• Sign Up

For verified travel tips and real support, visit: https://hopzone.eu/