Extending single-minus amplitudes to gravitons(openai.com) |
Extending single-minus amplitudes to gravitons(openai.com) |
Literally any out-of-distribution project in which I used LLMs lead to catastrophic failure. The models can't "see" stuff outside their training data.
You should already know what to ask to extract the answer OpenAI claims gpt-5.2-pro gave them.
Then you should be lucky to get an answer that makes sense.
Then you should already know how to verify the model's response.
Only after all these steps should you cherry-pick the one-in-a-million successful response to feature on your website.
And finally, you should prove that the answer didn't already exist in the training data. It's highly likely that the problem was solved before and the model picked that up. I have yet to see a genuinely novel discovery these models can produce.
* I'm an LLM researcher, but that doesn't mean I should close my eyes to the unjustified hype around language models.
The human authors have positions at the Institute for Advanced Study (Einstein's old institution), Vanderbilt, Harvard (Strominger) and Cambridge in the UK.
If you have to gauge this by the reputation of the experts involved in it as I do, that seems like a good list to me.
I don't see any reason to believe that this exact problem was solved before in the training data, but it's definitely an incremental result based on a very similar problem that the model had seen before.
For example if you have two gluons, you apply the rules of sum of spin and get
1 × 1 = 2 + 1 + 0
But for gravitons, the rules are
2 × 2 = 4 + 3 + 2 + 1 + 0
If you want to make give physicist nightmares and make mathematician cry, a tiny part of the details are in https://en.wikipedia.org/wiki/Table_of_Clebsch%E2%80%93Gorda... and https://en.wikipedia.org/wiki/Table_of_Clebsch%E2%80%93Gorda...
In conclusion, I'm not sure how difficult is to do the conversion from gluons to gravitons, but I'd recommend to run away.
[1] Assuming they exist.
If you can demonstrate that, I would put it to Strominger and his colleagues, and I imagine they would be obligated to cite your contribution in the peer-reviewed publication.
There's one little problem: OpenAI isn't actually open and doesn't reveal which dataset they used for training.