Construct · 1 fields · 1 citations

weak-to-strong generalization: a standalone construct

One name. One construct. A landmark on the map.

● anchor · a standalone construct Nearest construct by meaning·not indexed

A single standalone construct; nearby constructs are located against it in the Collider. Tap a field for its definition, unit, and measures.

AI safety

a strong model fine-tuned on labels from a weaker supervisor outperforms that supervisor, used as an analogy for humans supervising superhuman AI.

In AI safety, weak-to-strong generalization refers to a strong model fine-tuned on labels from a weaker supervisor outperforms that supervisor, used as an analogy for humans supervising superhuman AI. This meaning comes from Burns (2023) in arXiv.[1]

Unit
Recovery means
Measured by

Run it through the Collider, or take the citations with you.

See how weak-to-strong generalization sits against neighbouring constructs in the Collider, or export every reference on this page in one click.

Open in Collider Export BibTeX Report this entry

Nearby constructs

analogy4 fields · 4 citationsjingle · one word, four constructs generalization6 fields · 16 citationsjingle · one name, six constructs model3 fields · 3 citationsjingle · one word, three constructs

References

Burns, C., Izmailov, P., Kirchner, J. H., Baker, B., Gao, L., Aschenbrenner, L., Chen, Y., Ecoffet, A., Joglekar, M., Leike, J., Sutskever, I., & Wu, J. (2023). Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision. arXiv. https://doi.org/10.48550/arxiv.2312.09390

Cited on this page?

Confirm your construct, or correct us.

If your work is cited behind one of these entries, you know better than we do whether we captured the meaning right. Confirm it, suggest an edit, or dispute it, and your name goes on the entry as an author who has checked it, linked to your ORCID.

Claim your construct