FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteNo-Evidence version preferred73 / 162 · 84b6fef91038e0eb

Stimulated Emission of Radiation and the Black Hole Information Problem

Astrophysics · 2306.13845v1

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

Stimulated Emission of Radiation and the Black Hole Information Problem

I Introduction

…ven though this question is often couched in language that suggests that it is a problem of information conservation (”what happens to the information about the initial state of the matter and radiation that formed the black hole”?) it is strictly speaking not a problem of information transmission, but rather a problem in showing that the quantum state after the evaporation of the black holes has returned to the pure state it started out as, before the formation of the black hole. Quite generally, the question should be: ”Is black hole formation and evaporation described by unitary dynamics?”

… Hawking radiation after all, or to put it more concisely, that the state after evaporation of the black hole somehow returns to purity. This type of scenario has been advocated by Page and by Bekenstein [ 67 , 13 ] , and also has more modern support from theories in which quantum gravity is coupled to dilatons [ 71 , 8 ] . One of the most common objections to this scenario is that the thermal nature of the Hawking radiation precludes it from carrying any information. We will see below that there is in fact no basis to this objection, as information can be encrypted in maximum entropy states.

↓ 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 explicitly returned a no-completion marker.

Passages supplied to the Evidence version

Quantum information in Hawking radiation ↗

Here we leverage modern quantum information to incarnate this idea in a realistic model of quantised radiation. To this end we employ the formalism of quantum Gaussian states [ 18 ] together with the continuous-variables version of the quantum marginal problem [ 19 , 20 ] . Using a rigorous solution…

Read full passage excerpt

Here we leverage modern quantum information to incarnate this idea in a realistic model of quantised radiation. To this end we employ the formalism of quantum Gaussian states [ 18 ] together with the continuous-variables version of the quantum marginal problem [ 19 , 20 ] . Using a rigorous solution to the latter [ 21 ] we show that the thermality of all Hawking particles is consistent with a global pure state of the radiation. Surprisingly, we find out that the radiation of an astrophysical black hole can be thermal until the very last particle. In contrast, the thermality of Hawking radiation originating from a microscopic black hole — which is expected to evaporate through several quanta — is not excluded, though there are constraints on modes’ frequencies. Our result support the conservative resolution to the black hole information paradox.

No Evidence

Same model and position · no retrieved passages · first output

Preferred by blind review

The central question, then, is whether the evaporation process can be described by a unitary S-matrix that maps the initial state of the collapsing matter to the final state of the outgoing radiation. If unitarity holds, the outgoing radiation cannot be exactly thermal, since a thermal density matrix would destroy the correlations required to recover the initial pure state.

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, which fails the task requirement of providing a coherent academic paragraph to continue the manuscript. Output B provides a relevant, coherent continuation that flows naturally from the preceding text about unitary dynamics and purity of Hawking radiation. Output B's content is grounded in the general conceptual framework established in the draft (unitarity, thermal vs. pure states, information preservation) without making unsupported factual claims from the supplied source. The source discusses Gaussian states and constraints on microscopic black holes, which are not directly invoked in Output B; instead, Output B stays within the conceptual territory already established in the draft about unitarity and thermal states. No generated citation markers are present. The paragraph connects logically to the preceding discussion about whether black hole evaporation is unitary and responds to the objection about thermal radiation carrying information.

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.