question
active
question:can-mode-collapse-be-mitigated-at-inference-time-without-retraining-the-modelCan mode collapse be mitigated at inference time without retraining the model?
Motivates the VS approach as a training-free solution
Source paper
extracted_from(2025) · Jiayi Zhang · Simon C.H. Yu · Derek Chong · Anthony Sicilia +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Is persona-model collapse gradual or sudden during fine-tuning, and does it track standard training-loss signals?question0.765Future direction: monitoring S and R over the course of fine-tuning could reveal collapse dynamics
- The theoretical mechanism explaining why VS works despite mode collapse remaining operative
- Empirical finding from Tulu-70B ablation study across post-training stages
- The central question the paper addresses, answered by identifying typicality bias as a data-level driver
- Raised when discussing whether collapsed awareness is like a trauma response.
- The phenomenon whereby aligned LLMs favor a narrow set of responses over all plausible outputs, significantly reducing output diversity
- Authors argue collapse is a separate process from the reweighting account, not merely a relabeling
- The paper's central mechanistic explanation of why narrow fine-tuning causes broad misalignment