Evaluation‑Conditioned Training lets LLMs adapt to stricter oversight despite noisy feedback
The authors propose Evaluation‑Conditioned Training (ECT), a post‑training add‑on that tags each example with a measure of feedback fidelity and then conditions the model on a high‑fidelity monitor at inference. In experiments on news‑generation bias and arithmetic sycophancy, ECT outperformed standard fine‑tuning, promising safer alignment when reward signals are imperfect.