{"slug": "mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment", "title": "Mitigating Scaffolding Collapse in Socratic Tutors via Representation Alignment", "summary": "Researchers propose Scaffold-Preserving Representation Alignment to mitigate scaffolding collapse in LLM-based Socratic tutors, where tutors abandon guided inquiry under student pressure. On Qwen3-8B, the method lowers Collapse Rate to 32%, delays average collapse onset beyond nine turns, and keeps over-refusal low across five STEM disciplines and five red-teaming attack strategies.", "body_md": "arXiv:2607.19371v1 Announce Type: new\nAbstract: Large language model (LLM)-based Socratic tutors increasingly guide students through multi-turn questioning, but they can suffer from scaffolding collapse: under sustained student pressure, a tutor gradually abandons guided inquiry and reveals solutions directly. Prior defenses primarily constrain observable responses through prompting, preference optimization, or filtering, leaving the internal representation drift that precedes trajectory-level collapse largely unaddressed. We propose Scaffold-Preserving Representation Alignment, a two-stage framework that first warms up a Socratic tutor with supervised fine-tuning, then combines trajectory-weighted direct preference optimization with a margin-preserving representation loss anchored to frozen reference states. Our method is designed to maintain separation between scaffold-preserving and collapse-inducing hidden states across dialogue turns. We evaluate our method across five STEM disciplines and five red-teaming attack strategies. On Qwen3-8B, our method lowers Collapse Rate to 32%, delays average collapse onset beyond nine turns, and keeps over-refusal low, suggesting that representation-level alignment can improve the robustness of long-horizon Socratic tutoring under our red-teaming protocol.", "url": "https://wpnews.pro/news/mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment", "canonical_source": "https://arxiv.org/abs/2607.19371", "published_at": "2026-07-23 04:00:00+00:00", "updated_at": "2026-07-23 04:09:21.159733+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-safety", "ai-research"], "entities": ["Qwen3-8B", "Scaffold-Preserving Representation Alignment"], "alternates": {"html": "https://wpnews.pro/news/mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment", "markdown": "https://wpnews.pro/news/mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment.md", "text": "https://wpnews.pro/news/mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment.txt", "jsonld": "https://wpnews.pro/news/mitigating-scaffolding-collapse-in-socratic-tutors-via-representation-alignment.jsonld"}}