- Rating
- 3(+1)
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
The paper extends an existing compression-based elasticity model by incorporating dataset entropy and internal representation dimension into the relative weights of pretraining and alignment data. It recommends lower-dimensional alignment datasets or representations to improve robustness to subsequent fine-tuning. Three DistilGPT2/IMDb experiments compare initial positive-sentiment alignment and its retention under negative-review fine-tuning, using different data subsets and alignment methods.
Critical strengths
- The paper addresses a meaningful question: why equally sized alignment corpora or different alignment methods can produce different retention under subsequent fine-tuning. Distinguishing dataset structure from internal representation structure provides a useful perspective.
- The paper extends elasticity theory to investigate how data and representation structure may affect alignment rebound. The accompanying experiments provide preliminary evidence of retention differences across datasets and alignment methods beyond sample count.
Critical weaknesses
- It is unclear why lower dimension should reduce . Section 3.4 says lower-dimensional alignment data decrease . If lower dimension reduces alignment entropy , as the leading term in Eqs. (7) and (9) suggests, Eq. (3) instead gives a larger , with other terms fixed. Lowering internal dimension does the same for . The paper needs to explain how these relations support using lower-dimensional data or representations to reduce rebound.
- The experiments do not yet support the broader claims. All experiments use DistilGPT2 on IMDb sentiment continuation, so it is unclear whether the findings generalize to other datasets or alignment tasks. More importantly, the paper defines rebound as movement toward the pretraining distribution, while the evaluation relies on positive-output rates. These capture sentiment retention but do not directly measure that return. A metric more closely tied to the pretraining distribution would provide stronger support for the paper's interpretation.
Questions for authors
- Why should lower dimension reduce in Section 3.4, when reducing or in Eq. (3) increases it with other terms fixed?
- Can the authors assess rebound with a metric more directly tied to the pretraining distribution, and check whether the findings hold beyond IMDb sentiment continuation?
5 comments
We kindly remind you to read our rebuttal. We sincerely apologize for a typo in Line 303: it is an easily fixable typo by replacing decreases with increase, and all the other parts are based on the correct statement. So we hope you read our rebuttal and reconsider your score. Thank you very much!
We kindly remind you that the rebuttal period is drawing to a close. We sincerely hope to clarify the misunderstandings in our paper and address your concerns, and we truly appreciate your taking the time to read our response.
Response to Reviewer GKg8
We supplement our study with additional experiments across varying model scales, tasks, and random seeds. The results validate the conclusion that lower dimension alleviates alignment elasticity. Please refer to Q2a and Q2b of the General response for further details.
We conduct supplementary experiments to measure the KL divergence between the current model and the pretrained model during negative sentiment fine-tuning. The results show that the KL divergence between the two exhibits a decreasing trend, whereas it shows an increasing trend under direct fine-tuning without prior alignment. This indicates that the model exhibits a tendency to return toward the pretraining distribution.
Table 1
We sincerely apologize for the confusion this caused. This is a typographical error in our writing(Section 3.4). For detailed explanation and why dimension reduction weakens alignment elasticity, please refer to Q1 of the General response.
By measuring the KL divergence between the fine-tuned model and the pretrained model, the experiments (W2-2) confirm that the model exhibits a tendency to return toward the pretraining distribution.
We supplement our study with additional experiments across varying model scales, tasks, and random seeds. The results validate the conclusion that lower dimension alleviates alignment elasticity. Please refer to Q2 of the General response for further details.
We hope this clarifies our contribution and would appreciate your reconsideration of the score.