docs: quickstart says what E6 measured, not what it felt like - #5
Merged
Merged
Conversation
The quickstart said the return climbs out of the -300s within a few minutes. It is true in the weakest sense and useless in every other: the curve is noisy enough early that a seed is still dipping into the -300s at step 20k, and someone who watches for three minutes and sees noise concludes the thing does not learn. Now it says what the three seeds did - across zero at about 60k steps, 1928 +/- 224 at 500k, about a hundred minutes on the CPU-only machine that measured it - tells the reader the first few minutes are noise, and links the run. Both languages. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The quickstart said episode return "climbs out of the -300s within a few minutes". True in the weakest sense and useless in every other, because the early curve is noisy enough that a seed is still dipping into the -300s at step 20k. Someone who watches for three minutes and sees noise concludes the thing does not learn.
It now states the measured curve: across zero at about 60k steps, 1928 ± 224 at 500k, roughly a hundred minutes on the CPU-only machine that measured it — and says outright that the first few minutes are noise. Links the run.
Both English and Chinese pages.
mkdocs build --strictpasses.Numbers from
experiments/e6-first-learning-curve/summary.tsvinplugrl-server.🤖 Generated with Claude Code