Abstract
Long chain-of-thought (CoT) traces impose substantial output-token costs. Under constrained budgets, compression must preserve answer-critical information, making boundary placement central. Token-level and fixed-length boundaries can fragment coherent spans such as phrases, formulas, and local derivations, whereas step-level boundaries can bind content requiring different compression actions. We introduce SynLat, a text-latent CoT framework that aligns compression boundaries with syntactic structure through non-overlapping Syntax-Aligned Units (SAUs). An answer-conditioned Teacher constructs progressive KEEP/LATENT targets for a single compression-conditioned Student, which generates mixed reasoning from only the question and requested compression level at inference. Across two Qwen3 Student scales, Standard-CoT and Long-CoT groups, and three compression levels, SynLat matches or exceeds the strongest evaluated baseline in all 12 task-group aggregates and strictly leads in 11 under the reported achieved-CR selection protocol. Overall gains reach 3.6/2.6 points at MEDIUM and 7.0/5.5 points at HIGH for Qwen3-8B/14B, with larger advantages under stronger compression, particularly on Long-CoT groups.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Zhao, Y., Yu, H., Wang, S., Zhou, Y., Zhu, Z., Ning, Z., Mao, K., & Qian, J. (2026). SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning. https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning
MLA 9
Zhao, Yifeng, et al. "SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning." https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning.
Chicago (author–date)
Zhao, Yifeng, Hongjun Yu, Shibo Wang, Yunjiao Zhou, Zixiao Zhu, Zhipeng Ning, Kezhi Mao, and Junlang Qian. 2026. "SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning." https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning.
Harvard
Zhao, Y., Yu, H., Wang, S., Zhou, Y., Zhu, Z., Ning, Z., Mao, K. and Qian, J. (2026) 'SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning', Available at: https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning.
Vancouver
Zhao Y, Yu H, Wang S, Zhou Y, Zhu Z, Ning Z, et al. SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning. https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning
IEEE
Y. Zhao, H. Yu, S. Wang, Y. Zhou, Z. Zhu, Z. Ning, K. Mao, and J. Qian, "SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning," https://omanscience.com/en/articles/synlat-syntax-aligned-text-latent-compression-for-chain-of-thought-reasoning.