- Rating
- 3
- Confidence
- 3
- Informativeness
- 3
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
The paper proposes a data-efficient approach to train deep research agents. The authors generate challenging multi-hop questions and filter training trajectories based on maximum per-turn reasoning length as a proxy for looping reasoning. Experiments show that with only 2k trajectories and SFT alone, the resulting 35B model achieves strong performance across 6 benchmarks.
Critical strengths
-
The overall performance improvements are quite substantial, especially considering that the model is trained on only 2k trajectories.
-
The paper provides some interesting and useful insights into training data quality, for example, incorrect trajectories do not necessarily hurt performance, while trajectories with looping reasoning are harmful. The "less is more" finding is also interesting.
-
The authors conducted a fairly extensive set of experiments, including comparisons across multiple benchmarks, different model sizes, and several ablation studies.
Critical weaknesses
-
My main concern/question is that the contribution of discard-all context management is quite substantial. I'm not entirely sure how this relates to the main claim of data efficiency, since discard-all is an inference-time strategy adopted from prior work. While the experiments demonstrate the effectiveness of the training data, a large part of the reported performance also seems to depend on this additional inference-time strategy.
-
The maximum per-turn reasoning length seems to be a coarse proxy for looping reasoning, as in that it might discard some genuinely difficult trajectories that require long reasoning without being repetitive. I wonder how many such trajectories are filtered out. The paper also doesn't clearly specify how the filtering threshold is chosen.
-
This might be minor, but although the final training set contains only 2k trajectories, generating these trajectories could still be quite expensive, especially given the large number of tool calls involved. It would be helpful to report the total cost of generating these 2k trajectories, including the discarded ones, to better understand the end-to-end data efficiency.
Questions for authors
-
Have the authors looked into the cases that triggered discard-all context management? I'm curious whether these cases also contain looping reasoning. If so, simply excluding looping trajectories from SFT doesn't explicitly teach the model how to recognize or recover from such situations. It would be interesting to understand whether the loop-aware filtering actually reduces looping during inference, or whether discard-all is still needed to handle this failure mode.
-
How is the threshold for loop-aware filtering selected?
Response to Reviewer wUA8 (1/2): Entity Scaling and Task Difficulty
Dear Reviewer wUA8,
Thank you for your thoughtful questions about task difficulty and constraint redundancy. Below, we provide a controlled empirical comparison and a model of how entity scaling increases evidence-integration demands.
Empirical evidence. Appendix J compares low-hop tasks with 3–5 obfuscated entities along the root path and high-hop tasks with 7–9, using the same teacher and trajectory generator. We summarize the tool-call statistics and corresponding teacher accuracy below.
S1-DR*denotes S1-DR questions with trajectories regenerated by the same teacher.Increasing the number of hidden entities raises average tool calls by 3.82× and lowers teacher accuracy by 17.9 percentage points. Both measurements demonstrate greater retrieval and reasoning demands.
We agree with Q1 that relations in the question graph can be redundant, here we also present a theoretical model shows that increasing the number of hidden entities still makes synthesized questions harder by reducing the probability of a correct answer.
Let n count hidden entities, Ci the clues for entity i, C all clues, and A(S) candidate answers from a Wikidata query using clues S. Given ∣A(C)∣=1, define k=∑i=1n1[∣A(C∖Ci)∣>1].
k counts indispensable clue groups. For example, n=8 and k=3 indicate three indispensable groups among eight hidden entities. Our construction preserves a linear backbone, excluding diamonds or other complex topologies that introduce shortcuts. On this backbone, we model each added entity as independently adding an indispensable group with fixed probability α>0, giving E[k∣n]=αn.
With one correct candidate, d distractors, and score advantage Δ, softmax selection gives p(d,Δ)=eΔ+deΔ=1+de−Δ1.
Let L be context length in thousands of tokens and ρ the confusable-candidate density per thousand tokens, so d(L)≈ρL. The solution path integrates all k indispensable groups; applying the chain rule with a common conditional selection probability yields
Pground(k,L)≈(1+ρLe−Δ)−k.
Each additional indispensable group multiplies success by a factor below one. Independent backbone extensions therefore lower expected answer accuracy as the number of hidden entities grows. At 0.9 success per group, success falls from 72.9% for three groups to 43.0% for eight. This explains the observed increase in task difficulty.