FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteNo-Evidence version preferred38 / 162 · 4b033db27f1d104f

White Paper on Light Sterile Neutrino Searches and Related Phenomenology

Astrophysics · 2203.07323v3

FLOWING EVIDENCE BENCHMARK

How do we tell whether Evidence helps?

At the same writing position, with the same model and task, how does supplying retrieved paper passages change the first output? We compare matched versions and retain ties, unusable outputs and incomplete reviews.

SAME DRAFT · EVIDENCE ON OR OFF

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED INPUTS

SHARED BY BOTH

Manuscript context, model, task and prompt

A · With Evidence

Retrieved paper passages supplied

B · Without Evidence

No retrieved passages supplied

LLM

Same model and version

A → first output

B → first output

Blind judge agent

First outputs are anonymized as X and Y

Continuation review: accuracy, fit to the writing task and usability
Returns: X preferred / tie / Y preferred / both unusable

Order check: X / Y → Y / X

HOW DOES BLIND JUDGING WORK?

① Anonymize both outputs
The judge sees the same draft and both first outputs without knowing which received Evidence.

② Compare and swap order
The judge applies task-specific criteria in X/Y and then Y/X order.

③ Review disagreements
A third pass resolves disagreements. Incomplete reviews remain in the denominator.

How are autocomplete positions stratified?

Before seeing generation outcomes, we check whether a retrieved passage contains a specific proposition that directly supports the next writing move. Those positions appear in the left opportunity group; the rest are ordinary positions on the right. We select a balanced sample from admitted papers in each field. The 50/50 split is experimental, not a measure of how often either type occurs in writing.

AUTOCOMPLETE · 110

Biology, statistics and astrophysics

55 positions on each side; statistics uses the ten-paper rerun.

AUTOCOMPLETE · 52

Psychology and climate science

26 positions on each side; psychology includes nine papers and climate science four.

How are the table percentages calculated?

Across five fields there are 81 source-opportunity positions. The original blind review preferred Evidence in 51; a task check moved one empty Evidence continuation to both unusable, leaving 50 in public counts. In the ordinary group, another pair of empty outputs moved from no winner to both unusable. Original verdicts remain visible on case pages.

50Evidence version preferred
÷
81All positions in this group
=
62%Evidence preference in this group

Source contribution is a separate review: 16 of 21 Evidence wins entered into source review directly used retrieved papers; another 29 wins await review.

These are development-stage model judgments pending independent human review. They are not formal Benchmark conclusions and do not, on their own, establish causality.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

White Paper on Light Sterile Neutrino Searches and Related Phenomenology

Executive Summary > Primary Focuses for the Next Decade

Beyond direct anomaly tests, many alternate techniques/facilities, including direct neutrino mass measurements, long-baseline oscillation experiments, and atmospheric and astrophysical neutrino experiments, have been applied to the sterile neutrino explanation of the anomalies.

…he past ten years as viable interpretations of the experimental anomalies in the neutrino sector. The goal of these collective efforts will be to validate and solidify our understanding of the neutrino sector. Regardless of what the ongoing and upcoming experiments observe — be it a deviation from the three-neutrino picture or otherwise — the community should be prepared to address how to put these anomalies to test or adequately distinguish between different interpretations. We see the main experimental, analysis, and theory-driven thrusts that will be essential to achieving this goal being:

↓ Next paragraph was generated here

The highlighted text is near the writing position. This is a layout of the evaluation input, not a PDF screenshot.

A matched comparison at the same writing positionJudgment: No-Evidence version preferred

Both arms use the same model, task and manuscript context. Only the retrieved passages differ. The first recorded output from each arm is shown below.

Evidence

Same model and position · retrieved passages supplied · first output

Not preferred in this comparison

The product produced no insertable continuation text.

The model returned an empty text field.

Passages supplied to the Evidence version

A Model of Neutrino Anomalies and IceCube data ↗

The goal of this paper is to interpret both SBL neutrino anomalies and astronomical neutrino data observed at IceCube 4 4 4 However, in this paper we will not study phenomenological interpretations on astronomical-scale baseline IceCube data. See Refs. [ 10 , 9 ] . in terms of neutrino oscillations…

Read full passage excerpt

The goal of this paper is to interpret both SBL neutrino anomalies and astronomical neutrino data observed at IceCube 4 4 4 However, in this paper we will not study phenomenological interpretations on astronomical-scale baseline IceCube data. See Refs. [ 10 , 9 ] . in terms of neutrino oscillations in a convincing theoretical model where three light neutrinos become Pseudo-Dirac particles by introducing sterile neutrinos. Our model is different from the so-called 3 + n 3+n model with n n sterile neutrinos suggested to interpret SBL anomalies in terms of neutrino oscillations.

No Evidence

Same model and position · no retrieved passages · first output

Preferred by blind review

These thrusts span the direct detection of sterile neutrinos through anomalous spectral features, precision tests of the three-neutrino paradigm at short and long baselines, and the interpretation of anomalous excesses in light of new experimental and theoretical developments.

Judgment and task check

The blinded model review preferred the no-Evidence version. Its original rationale appears below.

Read the original model judgment

In this rationale, Output A = Evidence; Output B = No Evidence.

Output A is empty despite the task clearly requiring content to continue the manuscript paragraph. The draft sets up a colon introducing 'the main experimental, analysis, and theory-driven thrusts,' which demands a following list or elaboration. Output B provides a coherent academic continuation that enumerates these thrusts (direct detection, precision tests, interpretation of excesses) in a way that naturally follows from the draft's structure and topic. The content in B is general enough to be plausibly grounded in the draft's established scope (sterile neutrino searches, three-neutrino paradigm testing, anomaly interpretation) without requiring specific source support beyond what the draft itself establishes. No specific factual claims in B exceed the draft's framing.

The passages are input material; their presence does not mean an output used them or that they support every claim. Draft and source passages are excerpted; no simulated scores or PDF appear here.