Your submission is now in Draft mode.

Once it's ready, please submit your draft for review by our team of Community Moderators. Thank you!

Your essay is now in Draft mode

Once you submit your essay, it will be available to judges for review and you can no longer edit it. Please make sure to review eligibility criteria before submitting. Thank you!

Submit Essay

Once you submit your essay, you can no longer edit it.

Pending

This content now needs to be approved by community moderators.

Submitted

This essay was submitted and is waiting for review by judges.

GPT-3 scaling plateau in < 3 OOMs

Question

Consider figure 15 from this paper.

Some people (arguably the authors of this paper) predict that as we scale models past GPT-3's size (the 10^11 parameter learning curve, models with parameter count X trained on X elapsed tokens will score close to the L(D) line at X elapsed tokens.

We are interested in whether instead the trendline will "plateau" or at least be substantially slower than the line L(D) by the end of the next 3 orders of magnitude of parameter count. For the sake of specificity, let's say substantially slower = less than half as steep as L(D) on this graph.

If and when this graph is extended to 10^14 parameter models trained on 10^14 elapsed tokens of similar-quality data, will the 10^14 parameter learning curve have slowed down substantially?

This question resolves positively if the relevant experiment is done and reported (extending this graph, or providing equivalent data) and the slope of the learning curve for the 10^14 parameter model around 10^14 data points (Say, from 10^12 to 10^14) is less than half as steep as the slope of L(D). It resolves negatively if instead the slope is at least half as steep as L(D).

This question also resolves positively (or negatively) if it becomes uncontroversial what would have happened if the experiment had been done. For example, maybe other experiments will provide much more evidence about neural net scaling trends in general, such that it will be easy to calculate what would happen with this one in particular.

This question resolves ambiguously if by 2050 no one has performed the experiment AND there is no consensus about what would have happened if someone had.

The Metaculus moderators are the judge of final resort for determining whether and how to resolve this question.

For more context, see the comment threads below (Search for "It's a big deal.")

It is important that the data used to extend the graph be of similar quality. Obviously if we just threw in 10^14 tokens of basic arithmetic problems, the model would get good at basic arithmetic but not at anything else, and it's unclear whether the result would be on-trend or not. Ideally we'd have 10e14 tokens of diverse internet text, scanned books, chat logs, emails, etc. If this experiment gets done with different-quality data, the question becomes whether it gives us enough evidence to uncontroversial predict what would have happened if we had done it with similar-quality data.

Prediction

Note: this question resolved before its original close time. All of your predictions came after the resolution, so you did not gain (or lose) any points for it.

Note: this question resolved before its original close time. You earned points up until the question resolution, but not afterwards.

This question is not yet open for predictions.

Current points depend on your prediction, the community's prediction, and the result. Your total earned points are averaged over the lifetime of the question, so predict early to get as many points as possible! See the FAQ.

Metaculus help: Predicting

Predictions are the heart of Metaculus. Predicting is how you contribute to the wisdom of the crowd, and how you earn points and build up your personal Metaculus track record.

The basics of predicting are very simple: move the slider to best match the likelihood of the outcome, and click predict. You can predict as often as you want, and you're encouraged to change your mind when new information becomes available.

The displayed score is split into current points and total points. Current points show how much your prediction is worth now, whereas total points show the combined worth of all of your predictions over the lifetime of the question. The scoring details are available on the FAQ.

Thanks for predicting!

Your prediction has been recorded anonymously.

Want to track your predictions, earn points, and hone your forecasting skills? Create an account today!