Claims

A deterministic follow-up resolver raised intended-topic follow-up rates to 90–100% across five courses with all 400 scope checks passing, without changing the generator

Claim · All claims
Evidence · 1 study (3 entries) · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims

Reading the evidence codes

q is the design tier, i the impact, n the sample, and the kind with its r the rigour judged within that kind: what each code means.

This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.

q2 i? Topic checks rose from 23.3–40.0% in baseline resolver preflight to 90–100% live across five courses, with all 400 scope checks passing. → Sidney Shapiro and Joshua Lindemann 2026 q2 i? In the five-course run all 400 scope checks passed with zero API errors and each course reached the 90% follow-up target, while median per-turn latency ranged from 3.24 to 6.19 s. → Sidney Shapiro and Joshua Lindemann 2026 (2) q2 i? The 20 September router extension passed 306 focused Python tests and documented live probes on both nodes before activation. → Sidney Shapiro and Joshua Lindemann 2026 (3)

Evidence

Sidney Shapiro and Joshua Lindemann 2026

Sidney Shapiro and Joshua Lindemann. (2026). On-Premises Multi-Course RAG Tutoring for Business Education: Hardware–Software Trade-offs in a Campus AI Tutor. arXiv. https://arxiv.org/abs/2610.02510 doi:10.48550/arxiv.2610.02510

q2 · i? · design · r2

Multi-turn conversation evaluation with frozen fixtures of 20 dialogues and 80 turns per course; the decision summary reports topic checks rising "to 90–100% live across five courses" with the unchanged 8B generator.

"Topic checks rose from 23.3–40.0% in baseline resolver preflight to 90–100% live across five courses; all 400 scope checks passed."

Sidney Shapiro and Joshua Lindemann 2026 (2)

Sidney Shapiro and Joshua Lindemann. (2026). On-Premises Multi-Course RAG Tutoring for Business Education: Hardware–Software Trade-offs in a Campus AI Tutor. arXiv. https://arxiv.org/abs/2610.02510

q2 · i? · design · r2

The evaluation section reports the live five-course replay on Node B: "all 400 scope checks passed with zero API errors"; median per-turn latency ranged from 3.24 to 6.19 s, and Course 1's median rose from 4.92 to 5.81 s.

"In the five-course run, all 400 scope checks passed with zero API errors, and each course reached the 90% follow-up target."

Sidney Shapiro and Joshua Lindemann 2026 (3)

Sidney Shapiro and Joshua Lindemann. (2026). On-Premises Multi-Course RAG Tutoring for Business Education: Hardware–Software Trade-offs in a Campus AI Tutor. arXiv. https://arxiv.org/abs/2610.02510

q2 · i? · design · r2

The router extension evaluation used a frozen contract of 32 crossover cases plus focused checks; the section reports "306 focused Python tests and the documented live probes" passed before activation on both nodes.

"Both nodes passed 306 focused Python tests and the documented live probes before activation."

Discussion

Page metadata
Field Value
type claim
id deterministic-resolver-raises-topic-checks
title A deterministic follow-up resolver raised intended-topic follow-up rates to 90–100% across five courses with all 400 scope checks passing, without changing the generator
description A deterministic follow-up resolver raised intended-topic follow-up rates to 90–100% across five courses with all 400 scope checks passing, without changing the generator
status review
evidence_strength moderate
newest_evidence 2026
recheck_by 2029
generated by process:wiki-ingest on 2026-10-11
trust unverified — no one has confirmed this page's content
sources 3 — sidney-shapiro-and-joshua-lindemann-2026, sidney-shapiro-and-joshua-lindemann-2026-2, sidney-shapiro-and-joshua-lindemann-2026-3

Frontmatter is the page's machine-readable record — see the Schema & Guide.