works on
e-cusum 1language model decoding 1monitoring 1quantization 1reasoning models 1

From the 1 of 1 paper with an AI index.

researchers with a paper here
Showing cs.AIShow all

1 paper · 1 filter

cs.AI2026

Calibrated e-CUSUM Decoding for Quantized Reasoning Models: Why Token Log-Probability Is the Wrong Observable for Decoding Monitors

El Hassane Ettifouri, Ayoub Belfatmi, Mahaman Sanoussi Yahaya Alassan +1

The paper shows that using centered token log‑probability as a monitor for quantized reasoning language models is ineffective, and proposes a training‑free calibrated e‑CUSUM decod…