CreativityNeuro steers LLM weights to improve divergent thinking and cut mode collapse
ParshinShojaee · x · 2026-07-21
CreativityNeuro proposes steering language model weights rather than activations to improve divergent thinking and reduce mode collapse.
According to the paper, the method:
- uses contrastive creative prompts to identify creativity-relevant parameters
- applies a scaled weight perturbation to a sparse subset of those weights
- improves divergent-thinking behavior across multiple tasks
The authors say weight-space steering generalizes further out of distribution than activation-space steering.
More from Research
- Researcher flags unresolved data transparency issues in compositionality evals — najoungkim · 2026-10-27
- Separate Forms from Estimators: Why Steer by Difference of Means When You Can Use SGD? — johnhewtt · 2026-10-27
- EMNLP paper probes why some LLM biases survive rewording while others don't — radamihalcea · 2026-10-27
- Datology webinar: Curation Studio matched top data baselines with under 20% compute — josh_wills · 2026-10-27
- KANLab: an interactive browser playground for Kolmogorov-Arnold Networks — Lukas Niehaus · 2026-10-27
- CMU's LLM2Cert synthesizes stability certificates verified via SMT and Lean — CarnegieMellonU · 2026-10-27