Abstract
Long-horizon information-seeking agents often accumulate noisy or misleading context, causing early mistakes to persist and making recovery increasingly difficult. We introduce an autonomous search harness in which the agent manages its own search process through three states: Rubric, Answer, and Verify. The agent first defines criteria for a valid answer, searches under these criteria, and then independently verifies the result before deciding whether to terminate or continue searching. It is further equipped with a Seal Memory tool that enables active context management. Training this behavior with reinforcement learning, however, can induce Seal Collapse, resulting in unstable training and preventing the agent from reliably learning when and how to use its memory tools. We solve this with a simple strategy that trains only the final segment after context management. Our 35B model achieves 72.83 on BrowseComp, outperforming comparable open-source systems, and consistently improves over the base model across BrowseComp-ZH, xbench, DeepSearchQA, WideSearch, financial investigation, and product search. Ablations show that autonomous compression outperforms automatic compaction and validate our RL design.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Ma, J., Aster, L., Zhang, H., Song, S., Yuan, W., Zhang, Z., Jia, K., & Sui, Z. (2026). Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search. https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search
MLA 9
Ma, Jingyuan, et al. "Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search." https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search.
Chicago (author–date)
Ma, Jingyuan, Lynx Aster, He Zhang, Siyao Song, Weijie Yuan, Zhe Zhang, Kai Jia, and Zhifang Sui. 2026. "Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search." https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search.
Harvard
Ma, J., Aster, L., Zhang, H., Song, S., Yuan, W., Zhang, Z., Jia, K. and Sui, Z. (2026) 'Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search', Available at: https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search.
Vancouver
Ma J, Aster L, Zhang H, Song S, Yuan W, Zhang Z, et al. Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search. https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search
IEEE
J. Ma, L. Aster, H. Zhang, S. Song, W. Yuan, Z. Zhang, K. Jia, and Z. Sui, "Traverse: Learning When to Remember, Reset, and Redirect for Long-Horizon Web Search," https://omanscience.com/en/articles/traverse-learning-when-to-remember-reset-and-redirect-for-long-horizon-web-search.