A frozen language model on reasoning tasks has two coupled weaknesses: it under-uses evidence its own residual stream already encodes, and it fails to detect when the input is insufficient to answer, so it confabulates. This paper consolidates two research lines that address these on the same residual stream: a conditi...
Ziyang Luo, Zhong-Yao Chu, Xin-Jie He et al.· 0 citations
These exploratory offline results characterize ranking, probability quality and interface failures; they establish neither a causal typed-interface benefit nor general superiority, calibration or end-to-end acceleration.
Xu-Kui Qin, You Wang, Xin-Jie He et al.· 1 citation
Noise2Noise (N2N) trains denoisers on pairs of independently corrupted observations, eliminating clean references. We stress-test two natural conjectures about why the L1 loss outperforms L2 here. First, the hypothesis that the L1 loss confers robustness via parameter sparsity confuses the loss with Lasso regularizatio...
Ding-Yan Shang, Zhen-Yu Xu, You Wang et al.· 0 citations
Context masking is established as necessary for attributing early answer availability to an explanation rather than its hidden input when early-prefix evidence disappears after masking.
Bo-Nan Shen, Ding-Yan Shang, You Wang et al.· arXiv.org· 0 citations
The results support truncated CoT auditing as a lightweight process-level diagnostic for answer-driven reasoning in math tutoring explanations and support truncated CoT auditing as a lightweight process-level diagnostic for answer-driven reasoning in math tutoring explanations.
Bo-Nan Shen, Ding-Yan Shang, You Wang et al.· 0 citations
Naming the benchmark, metric, target behavior, and model panel is the minimum a safety claim needs, and both instruments score harmful compliance, so this is evidence of convergent validity rather than general safety.
You Wang, Xiao Han, Ding-Yan Shang et al.· arXiv.org· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
