Tech Meridian ← LIVE FEED
PROMY MERIDIAN RU

RESEARCH · RESEARCH · #3643

Paper operationalizes 'opaque serial depth' as NLS depth to quantify unverbalized serial cognition

The paper (arXiv:2611.17495v1) defines a concrete metric—NLS depth—based on "NL-rooted" nodes to proxy the length of a model's computation paths that bypass interpretable bottlenecks, and shows NLS depth can be adjudicated from architecture, training recipe, and deployment without post‑training probes. The authors report that NLS depth correlates with layer count across open models, would likely grow slowly under standard transformer scaling but could rise substantially for latent‑reasoning architectures, provide a worked bound for Gemma3‑1B, and recommend developers track and report NLS depth alongside CoT monitorability measures.

KEY POINTS

  1. The paper (arXiv:2611.17495v1) defines a concrete metric—NLS depth—based on "NL-rooted" nodes to proxy the length of a model's computation paths that bypass interpretable bottlenecks, and shows NLS depth can be adjudicated from architecture, training recipe, and deployment without post‑training probes.
  2. The authors report that NLS depth correlates with layer count across open models, would likely grow slowly under standard transformer scaling but could rise substantially for latent‑reasoning architectures, provide a worked bound for Gemma3‑1B, and recommend developers track and report NLS depth alongside CoT monitorability measures.
  3. Provides a pre‑deployment, architecture‑level metric to quantify how much unverbalized serial cognition a model can do, informing monitorability and oversight choices.

WHY IT MATTERS

Provides a pre‑deployment, architecture‑level metric to quantify how much unverbalized serial cognition a model can do, informing monitorability and oversight choices.

SOURCES & TIMELINE

1