Counterfactual Constraint-Conditioned On-Policy Distillation for Multi-Constraint Instruction Following
arXiv:2609.27421v1 Announce Type: cross Abstract: Multi-constraint instruction following requires a model to respond to a query under many simultaneously active constraints. Even strong instruction-tuned models still routinely violate some of them. Existing approaches either augment supervision wit…
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-09-24 04:00 · arXiv stat.ML
Counterfactual Constraint-Conditioned On-Policy Distillation for Multi-Constraint Instruction Following