Model size defines a core trade-off: too-large SLMs answer without the cognitive structure while too-small SLMs break instruction-following coherence¶

Claim · All claims
Evidence · 1 study · design r2 · q1
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q1 i? The authors report that if the model is too large, the SLM answers correctly on its own without needing the proposed cognitive structure, while if too small, its coherence in following rigid instructions breaks down, defining the core trade-off between explicit control and scalable coverage. → Hyeongdon Moon 2026

Evidence¶

Hyeongdon Moon 2026¶

Hyeongdon Moon, Carolyn Rosé, and John Stamper. (2026). Cognitive Agent Compilation for Explicit Problem Solver Modeling. https://arxiv.org/abs/2605.07040 doi:10.48550/arxiv.2605.07040

q1 · i? · design · r2

Authors' challenge analysis from the proof of concept (Section 6.1), describing the 'curse of prior knowledge'. They state this "defines the core trade-off in CAC between explicit control and scalable coverage". No effect sizes are printed.

"If the model is too large, the SLM answers correctly on its own without needing the proposed cognitive structure; conversely, if the model is too small, its coherence in following rigid instructions breaks down."

Discussion¶

Related Claims¶

• An SLM executor under tight constraints can drive a Cognitive Agent through formatted outputs and logit-based action selection, yielding partial rather than full explainability — related
• A 4-bit quantized 27B local Teacher LLM generated knowledge in the required format, suggesting accessible-resource implementations may be feasible
• MCQs generated by the generate-then-validate SLM pipeline are generally answerable and unambiguous, as judged by seven human experts
• At the implementation checkpoint, several groups added an explicit human-approval step after encountering local-LLM structured-output reliability limitations — related