- Rating
- 1
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 1
- Soundness
- 1
Read review
Paper summary
This paper evaluates the safety setup where the aim is to separate emergence violations from propagation. The paper shows that most violations appeared with exposure. It works with frontier models. It uses a few steps: a founder (G0), a successor (G1), and another successor (G2). The idea is to check the impact of earlier models on successor ones.
Critical strengths
This is an important question for evaluating the impact of earlier models on successor models.
Critical weaknesses
- This paper is hard to follow, and there are missing definitions and examples to follow the setup the paper uses. In addition, the paper can benefit from more intuition and more explanation of how to interpret the results.
- The paper introduces a specific setup to evaluate the impact of earlier models on successor models; to validate the results, ablation of the configuration this work uses is needed.
- The number of blocks used across the paper is low (in many results less then a 100) reducing our comprehending the results. In addition I have not found what the questions look like or which dataset they are from.
Questions for authors
- Line 130: What are a cohort and a successor branch? These definitions are important throughout the paper, and when they are first introduced, there is no explicit definition.
- What is invalid advice (Figure 1)?
- Line 158: What is “turnover”?
- Section 3 can benefit from a running example.
- Table 2 is referred to before Table 1; please fix the order of the tables.
- Please add examples to Table 2. This table is not intuitive, and it is hard to understand what each setup means.
- Line 188: “PRESSURE adds performance-review language favoring higher“ - What is this? Is this text something the model gets somehow in the prompt? What does this text look like?
- What datasets are used in Section 4?
- Please add intuition on how to read each figure and the results behind it.
- Figure 2,3 can benefit from improving their visuality, they are hard to read in their current format.