SOURCE-LINKED INTELLIGENCE

Off-Target Effects of Response-Style Alignment in a Korean 27B Language Model

arXiv · AI, language, vision and robotics · article · Sep 10, 2026 · UTC

We post-train Qwen3.8-27B for Korean response style -- verbosity, list and markdown usage, discourse structure and register -- and measure two behaviours the objective never targets: abstention on ambiguous social questions in KoBBQ, where the benchmark-correct answer is UNKNOWN, and unprompted disclosure in securities guidance. Both move, and the changes are expressed primarily through the model's emission policy: how often it answers and how much it says. Matched target-form controls show that answer propensity depends on the training target, not the prompt set or recipe alone. Holding promp

Read original source ↗ Open in workspace

recordType

paper

region

Global