- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3(+1)
- Soundness
- 2
Read review
Paper summary
The authors propose the SAVE framework, which samples multiple on-policy trajectories per question, partitions them into direct-answer and tool-assisted groups, and constructs a hierarchical trajectory reward comprising three binary components: task success, a cross-strategy contrastive signal (activated only when one strategy succeeds), and an intra-tool minimum-depth preference (favoring the shallowest successful tool-assisted trajectory). The policy is optimized via REINFORCE Leave-One-Out (RLOO). The training pipeline includes difficulty-aware RL instance selection from ChartQA and SFT initialization based on a teacher model (Qwen3.5-397B). The primary claim is that across all 15 backbone-benchmark pairs, SAVE improves Avg@3 by 8.33 percentage points on average over SFT, while reducing the macro-average number of tool calls per tool-using trajectory by 8.6%.
Critical strengths
- Separating the invocation decision (cross-strategy comparison) from the stopping decision (intra-tool minimum-depth preference) constitutes the core conceptual contribution of the paper. The paper clearly demonstrates why a unified per-call penalty invariably favors direct trajectories over tool-assisted ones, providing a principled motivation for SAVE's alternative. The binary non-redundancy metric (Equation 11) elegantly avoids the need to tune a per-call cost coefficient. The ablation study (Table 3) confirms that both components are necessary. The equal-weight design requires no reward hyperparameter tuning, which is favorable for practical adoption.
- The paper reports paired bootstrap confidence intervals for all 15 backbone-benchmark comparisons, honestly disclosing that 5 of these intervals contain zero. Controlled tool-access sensitivity analysis, one-step stopping diagnostics, and training-time reward diagnostics provide multiple complementary perspectives. This level of statistical rigor and honest reporting exceeds the norm in the visual tool-use RL literature.
Critical weaknesses
- The applicability of the method is limited. SAVE is evaluated exclusively with local cropping operations applied to static images, and all training data comes from ChartQA. Four of the six evaluation benchmarks involve chart or document understanding (CharXiv, ChartQAPro, HR-Bench-4K/8K), which are closely related to the training domain. The paper positions itself as a general framework for "adaptive visual evidence acquisition," but this overstates what is actually demonstrated.
- A controlled comparison with the most closely related concurrent methods is missing. Beacon (Wang et al., 2026c) conditions its adaptive reward on whether the rollout group contains a correct tool-free answer, addressing nearly the same problem as SAVE's cross-strategy contrast. It is discussed in related work but absent from Table 2. BEE and AdaTooler-V are similarly cited but not experimentally compared. A controlled experiment training competitive alternative rewards on the same SFT-initialized model and the same ChartQA data would more convincingly isolate the contribution of SAVE's hierarchical reward design.
Questions for authors
- How does SAVE compare experimentally with Beacon (Wang et al., 2026c), which also conditions rewards on observed strategy success within rollout groups?
- Training uses only ChartQA, yet the largest gains appear on HR-Bench-8K (+27.97 pp for Qwen2.5-VL) and RealWorldQA (+24.26 pp), which are far outside the chart domain. What explains this strong positive transfer?
- Given that 56.52% of RL training instances are "hard" (Table 5, n+(xi)=0), and all trajectories receive zero reward when no strategy succeeds, does this mean over half of training batches contribute no gradient signal from the hierarchical components? How does this affect learning efficiency?
5 comments
On Strong Cross-Domain Transfer (Q2)
3. On Strong Cross-Domain Transfer (Q2)
We attribute the positive transfer primarily to improved visual evidence acquisition behavior. Although training uses ChartQA, the learned behaviors of when to invoke tools, how to use local visual observations, and when to stop can transfer to other visual question-answering tasks.
Table R1. Cross-domain performance and tool-use behavior of Qwen2.5-VL-7B. Acc denotes Avg@3 (%), Rate denotes tool-use rate (%), and Calls denotes Calls/Used. Improvements are reported in percentage points.
| Benchmark | Method | Acc (%) | Rate (%) | Calls |
|---|---|---|---|---|
| HR-Bench-8K | SFT | 31.29 | 8.38 | 1.776 |
| HR-Bench-8K | SAVE | 59.26 | 20.25 | 1.609 |
| HR-Bench-8K | Improvement | +27.97 | +11.87 | -0.167 |
| RealWorldQA | SFT | 40.92 | 31.98 | 1.546 |
| RealWorldQA | SAVE | 65.18 | 48.24 | 1.290 |
| RealWorldQA | Improvement | +24.26 | +16.26 | -0.256 |
As shown in Table R1, both benchmarks exhibit substantial accuracy improvements, accompanied by higher tool-use rates and lower Calls/Used, indicating more frequent evidence acquisition with shallower tool-assisted trajectories.
We further conducted controlled tool-access experiments under matched evaluation settings.
Table R2. Controlled tool-access sensitivity analysis on Qwen2.5-VL-7B (Appendix G.2.1, Table 12). Gain denotes Auto-tool minus No-tool Avg@3 in percentage points, with paired 95% bootstrap confidence intervals.
| Benchmark | Checkpoint | No-tool Avg@3 (%) | Auto-tool Avg@3 (%) | Gain (pp) [95% CI] |
|---|---|---|---|---|
| HallusionBench | SFT | 51.408 | 64.884 | +13.476 [11.479, 15.437] |
| HallusionBench | SAVE | 57.077 | 67.201 | +10.125 [8.306, 11.907] |
| HR-Bench-8K | SFT | 26.583 | 29.417 | +2.833 [0.583, 5.042] |
| HR-Bench-8K | SAVE | 28.333 | 36.958 | +8.625 [6.542, 10.708] |
Table R2 shows that enabling autonomous tool access improves SAVE's performance on HR-Bench-8K by 8.625 percentage points, providing additional evidence that local visual evidence acquisition is beneficial on this benchmark.
We also examined whether an additional crop is beneficial after the model decides to stop or continue.
Table R3. One-step Stop/Continue diagnostic on HR-Bench-8K (Appendix G.2.2, Table 13). Stop evaluates immediate answering, while Continue-one forces one additional model-selected crop. The difference is Continue-one minus Stop.
| Checkpoint | Decision Group | Stop Avg@3 (%) | Continue-one Avg@3 (%) | Difference (pp) [95% CI] |
|---|---|---|---|---|
| SFT | Stop | 51.852 | 44.444 | -7.407 [-12.346, -2.469] |
| SFT | Continue | 46.032 | 46.561 | +0.529 [-7.407, 8.466] |
| SAVE | Stop | 45.528 | 37.398 | -8.130 [-15.447, -1.626] |
| SAVE | Continue | 43.478 | 47.826 | +4.348 [-5.797, 14.493] |
For histories where SAVE decides to stop, an additional crop reduces Avg@3 by 8.130 percentage points. Together, these findings suggest that the evidence acquisition and stopping behaviors learned from ChartQA transfer to other visual tasks, improving the model's ability to use local visual information effectively.
On Hard Training Instances and Learning Efficiency (Q3)
The 56.52% Hard instances are defined by the teacher model failing in all three sampled trajectories. This teacher-based difficulty classification is distinct from the actor's success during online reinforcement learning.
Table R4. Difficulty distribution of the curated RL training dataset (Appendix D, Table 6). Difficulty is determined by the number of correct trajectories among three teacher-generated samples.
| Difficulty | Criterion | Instances | Proportion (%) |
|---|---|---|---|
| Medium | At least one but fewer than three correct teacher trajectories | 3,395 | 43.48 |
| Hard | No correct teacher trajectories | 4,414 | 56.52 |
| Total RL data | Medium + Hard | 7,809 | 100.00 |
Appendix G.4 (Figure 8) provides training-time evidence of active hierarchical reward signals**. **Figures 8(a) and 8(d) show that rollout groups with successful strategies and cross-strategy contrasts occur throughout training, providing opportunities for strategy-level learning. Figure 8(b) shows the continued presence of successful tool trajectories at the minimum observed depth, while Figure 8(c) shows an overall reduction in redundant correct tool use, consistent with the intended minimum-depth preference.
Moreover, Figure 3 demonstrates improvements in validation accuracy and tool-use behavior, indicating effective policy learning during training.
For groups in which all trajectories fail or eligible rewards are identical, Appendix E.1.1 explains that rejection sampling filters these groups, and additional groups are sampled to construct optimization batches with informative reward differences. The associated efficiency cost mainly comes from additional rollout generation and answer evaluation.
We thank the reviewer again for these valuable suggestions, which have helped us strengthen the experimental analysis and clarify the scope of our method.
On the Applicability of SAVE (W1) & On Comparisons with Related Methods (W2 / Q1)
We sincerely thank the reviewer for the constructive comments. We have updated the latest version of the manuscript to clarify the scope of SAVE and added comparisons with related methods, cross-domain transfer analyses, and training-time diagnostics.
1. On the Applicability of SAVE (W1)
We have revised the manuscript to clarify that SAVE focuses on adaptive visual evidence acquisition through local cropping for static-image question answering. Our experiments use ChartQA for training and evaluate the method on six image-question answering benchmarks. We have refined the positioning of SAVE to align with the evaluated setting.
2. On Comparisons with Related Methods (W2 / Q1)
We added experimental comparisons with Beacon, BEE, and AdaTooler-V in Table 2. For Beacon, we used its officially released Qwen3-VL-8B-Instruct checkpoint. For BEE and AdaTooler-V, we implemented the corresponding methods following their original papers. All results were obtained under a unified visual cropping interface and evaluation protocol.
Table 2. Comparison with BEE, AdaTooler-V, and Beacon. Acc / Rate / Calls represent Avg@3 (%), tool-use rate (%), and Calls/Used, respectively.
Although Beacon and SAVE both leverage observed strategy success within rollout groups, their reward preferences differ. When both Direct and Tool-assisted strategies succeed, Beacon downweights successful tool-assisted trajectories, whereas SAVE assigns equal rewards to successful Direct trajectories and minimum-depth successful Tool-assisted trajectories, while assigning lower rewards to longer successful Tool-assisted trajectories. This explicitly encourages adaptive stopping within the tool-assisted strategy.
As shown in Table 2, under the same Qwen3-VL-8B-Instruct backbone and cropping interface, SAVE outperforms Beacon on all six benchmarks, achieving a macro-average Avg@3 of 60.43% versus 57.48%, an improvement of 2.95 percentage points. SAVE also outperforms BEE and AdaTooler-V on all six benchmarks in the Qwen2.5-VL-7B comparison group.