I-D Action: draft-gaikwad-llm-fault-detection-methodology-00.txt

Internet-Draft draft-gaikwad-llm-fault-detection-methodology-00.txt is now
available.

   Title:   Benchmarking Methodology for Output Behavior Fault Detection in Large Language Model Serving Systems
   Author:  Madhava Gaikwad
   Name:    draft-gaikwad-llm-fault-detection-methodology-00.txt
   Pages:   13
   Dates:   2026-08-11

Abstract:

   This document defines test procedures for characterizing the fault
   detection capability of observability systems that monitor Large
   Language Model (LLM) serving deployments.  Procedures are given for
   Detection Latency, Detection Coverage, Detection Threshold Magnitude,
   False Detection Rate, and Boundary Masking.

   The Detector Under Test is the observability system.  Output Behavior
   Faults are injected at a known time under controlled conditions,
   which makes detection latency directly measurable.

   This document is a companion to "Benchmarking Terminology for Output
   Behavior Fault Detection in Large Language Model Serving Systems" and
   is to be read alongside it.  This document specifies no acceptance
   thresholds.

The IETF datatracker status page for this Internet-Draft is:
https://datatracker.ietf.org/doc/draft-gaikwad-llm-fault-detection-methodology/

There is also an HTML version available at:
https://www.ietf.org/archive/id/draft-gaikwad-llm-fault-detection-methodology-00.html

Internet-Drafts are also available by rsync at:
rsync.ietf.org::internet-drafts


_______________________________________________
I-D-Announce mailing list -- [email protected]
To unsubscribe send an email to [email protected]