All my GPT prompt templates have reset back to gpt-4o even though the template got saved as gpt-4o-mini Seems incredibly dangerous and expensive to reset everyone's model choice settings without telling them
We recently consolidated our AI experience across different providers (Claygent, Anthropic, OpenAI, and Gemini) into a single "Use AI" enrichment.When using the "Use AI" integration, you can now select between any OpenAI, Anthropic, or Gemini models for non-online cases, or choose Claygent models for web research.By default, the system uses our Clay-managed key, but you have complete control to select your preferred model and account: 1. Choose your model under "Model" settings 2. Select your account under "Account" settings 3. You can add your own API key through "Model > Account > Add Account" if you preferI recommend reviewing your model settings to ensure they align with your preferences. Sources: AI action consolidation Use AI - Clay University
Looks like you're still looking for help. We're escalating this over to our support team and someone should be in touch soon. In the meantime, feel free to include any other information that may be helpful in solving your issue!
Hey Chris! Let me help clarify this: I just ran a test to check this behavior. While it might look confusing, what I found is: 1. Even though area #2 shows "GPT 4o", if you saved your template with GPT 4o Mini, it should still actually run with Mini (that's what area #1 indicates) I tested this in this table and it worked as expected: https://app.clay.com/workspaces/10786/tables/t_6AT5J2uyYo5S/views/gv_qkJaXN72jQtP Could you confirm if you're seeing something different? Like, is the model in area #1 also not showing Mini? I'm going to flag this to our engineers based on your response Let me know and I'll help investigate! 🕵️♂️ https://downloads.intercomcdn.com/i/o/w28k1kwz/1265246191/f92168bb21db693fa67d497bd99c/CleanShot+2024-11-25+at+_41MZGMr0gp%402x.png?expires=1732563900&signature=08567fb9706d9c8eca1a47c0d3be40459188dbb9ebd95a7e6a29424e6d4dd255&req=dSIhE8t6m4BWWPMW1HO4zbTznUmzbgTmYQJo7MhAFRaDYAHf5zpJDUIvVFVs%0AgzOfxT3LcNELD6tiMMs%3D%0A
In my test, it was only using the GPT 40 Mini (the #1) not the (#2). I'll go ahead and let our team know that this is happening so we can work on fixing it! :)
Bo (. Thanks. But just to be clear, I would advocate for not changing it from the template. I think the model selector should come in as whatever the default model choice was when the template was saved, and not just default visually to gpt-4o. I'm terrified my team is going to run 50k rows with gpt-4o turned on accidentally
Totally get what you're saying - we're actually on the same page! The model shown in spot #2 should match whatever model was saved with the template (spot #1). They should be consistent since that's the model it's actually using, which was the model that was previously saved w
Sounds good - thanks for the help!
Hey — just wrapping this up to confirm the bug is no longer occurring and has been resolved. Can you let me know if everything looks good on your side? Let me know if you have more questions.
.png){kind=logo}
