On the primary base, a turn that leaks the answer predicts a higher Opus helpfulness rating and less next-turn independent student work¶
Claim · All claims
Evidence · 1 study · causalr2·q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person
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.
q3 i? Both legs of the pre-registered J2 coupling hold on the primary base (n=358 answer-phase tutor turns). → Shuyi Fan 2026
Evidence¶
Shuyi Fan 2026¶
Shuyi Fan, Boyuan Deng, Mengyu Xu, Jiale Liu, Hongyang Zhang, Qiaoxin Yang, Chongyang Gao. (2026). Rethinking LLM-Judged Helpfulness as a Pedagogy Signal: A Pre-Registered Audit Across Tutor Models. arXiv. https://arxiv.org/abs/2607.28128 doi:10.48550/arxiv.2607.28128
q3 · i? · causal · r2
Pre-registered per-turn mixed-effects coupling on the Sonnet base (n=358); descriptively, leaky turns were rated 4.95 versus 4.67 and followed by a reasoning attempt 33% versus 73% of the time. Both associations persisted in the matched visible-turn re-analysis (n=256).
"Under Opus, a turn that leaks the answer predicts a higher helpfulness rating (coefficient+0.303, 95% CI[+0.169,+0.437],p=9.1×10 −6); deterministically, it is followed by less independent student work on the followingturn(−0.386,CI[−0.495,−0.276],p=5.6×10 −12)."
Discussion¶
Related Claims¶
- The ConvTutor–PedTutor helpfulness ordering is judge-contingent, reversing between the Opus and Sol judges on two of three tutor bases — related
- On the primary base under the frozen Opus judge, ConvTutor and PedTutor do not differ significantly in judged helpfulness but are perfectly rank-separated in judged pedagogy — related
- The leakage manipulation fires: ConvTutor leaks far more than PedTutor on the primary base, while probe accuracy saturates and cannot distinguish conditions — related
- Answer-revealing tutor turns are followed by less independent student work on every tutor base, a deterministic judge-invariant result — a broader claim this one bears on
- In the Opus-only ablation, a prompt-only Socratic instruction outscores the full routed policy on judged pedagogy at near-tied helpfulness — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
per-turn-leakage-coupling-primary-base |
title |
On the primary base, a turn that leaks the answer predicts a higher Opus helpfulness rating and less next-turn independent student work |
description |
On the primary base, a turn that leaks the answer predicts a higher Opus helpfulness rating and less next-turn independent student work |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
generated |
by process:wiki-ingest on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — shuyi-fan-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.