In our engagements with governments, AI safety institutes, and frontier AI developers, we found the concept of the “evaluation gap” (short: ‘evals gap’) helpful to communicate the current state of the art and what is needed for the field to move towards more robust evaluations. In this post, we briefly explain the concept and its implications. For the purpose of this post, “evals” specifically refer to safety evaluations of frontier models.

Evals have become a prominent tool underpinning governance frameworks and AI safety mechanisms. Given that, we are concerned that policymakers and industry players both (i) overestimate the number of currently available high-quality evals and (ii) underestimate the time it takes to develop them. In our experience, available evals are not sufficient (in quality and quantity) to robustly identify the capabilities of existing and near-future models.

We call this overarching idea the evals gap. Unless more focused attention is [...]

---

Outline:

(01:37) Evaluations underpin many high-stakes decisions

(02:42) Current evaluations are insufficient to underpin high-stakes decisions

(02:48) Many important evals don’t yet exist

(04:06) Coverage is typically low

(05:42) Development and interpretation of evals is complicated

(08:12) Proper elicitation is an unsolved research question

(09:35) Without action, the evals gap may widen

(12:28) Closing the evals gap is possible

The original text contained 5 footnotes which were omitted from this narration.

The original text contained 1 image which was described by AI.

---

First published:

Source:

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.