- Rating
- 1
- Confidence
- 2
- Informativeness
- 1
- Presentation
- 1
- Soundness
- 2
Read review
Paper summary
This paper proposes BCAR, a training-freemethod that allocates a prompt budget across retrieved turns. It used query-dependent value and concave water-filling. On LongMemEval-S, it improves accuracy over retaining the full retrieved pool while reducing prompt tokens approximately eightfold. AgentLongBench shows smaller average gains, and latency measurements demonstrate reductions in TTFT.
Critical strengths
- The problem decomposition is helpful. Separating budget allocation across turns from evidence selection within turns provides a modular approach.
- BCAR beats both retrieve-all and several compression baselines at roughly comparable compressed prompt sizes.
Critical weaknesses
- As also ackowledged by the authors, the paper was drafted using AI. The result is that the paper is hard to read and follow, without a clear, professional academic writing sytles. For example, "Perceived latency (prep + TTFT)" is a confusing and not well-structured phrase to appear in abstract. Also, why are negative results and limitations highlighted in contributions ("neither pooled comparison is significant, and the two 64k knowledge-intensive splits are slightly negative")?
- Due to 1, the paper looks like a technical report instead of an academic paper and no insightful conclusions are drawn.
- The methodology is hard to follow due to 1 as well. It is unclear how the designs are motivated.
Questions for authors
Please see weaknesses.