FlowingFlowing
← Back to all development cases
This case was judged by an anonymized AI reviewer and has not been independently reviewed by people; inspect the draft, cards, and both outputs to assess it.
AutocompleteWith cards better195 / 280 · beu-11e752e54d4fc8f0

Non-singular cosmological scenarios in scalar-tensor theories and their stability: a review

Astrophysics · 2409.16108v2

FLOWING EVIDENCE BENCHMARK

How do we compare writing with and without cards?

At the same writing position, the same model produces one first result with cards and one without. All 280 pairs enter the overall anonymized review, which assesses whether a decision not to continue fits the draft.

Overall: 138 with-card wins, 26 ties, 107 no-card wins, 5 pairs where neither was usable, and 4 inconclusive judgments. Both sides wrote text in 218 pairs; 48 had a deliberate stop and 14 included an invalid output.

SAME DRAFT · WITH OR WITHOUT CARDS

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED VERSIONS

SHARED BY BOTH

Draft, writing task, and generation model

A · With cards

Relevant literature cards supplied

B · Without cards

No literature cards supplied

Generation model

Same model and version

A → first output

B → first output

Anonymized AI judge

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

Review again with the answer order swapped

HOW DOES BLIND JUDGING WORK?

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

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 positions and cards selected?

We selected 40 writing positions in each of seven fields, for 280 comparisons. The position and relevant literature cards were set before each comparison, so both versions began from the same draft.

The main result retains every frozen position, including places where cards might not add a useful next move. This avoids selecting cases based on their outputs, but 49% describes the whole cohort, not a win rate where evidence clearly had an opportunity to help. This cohort did not record that opportunity label for each position before generation, so those two groups cannot yet be separated reliably.

CARD USE · 140/280

Reported card use in continuations with text

The model reports whether a card affected its continuation. In 140 of 280 tests it produced continuation text and reported card use. The previous count of 150 included 10 cases without continuation text and has been corrected.

COMPARISON · 70 / 7 / 30

Which written version was better?

Of the 107 reported-use cases where both versions produced text, the blinded AI judge preferred the with-card version 70 times, found 7 about the same, and preferred the without-card version 30 times.

How are the two table percentages calculated?

ADOPTION COVERAGE

140reported use with text
÷
280all positions
=
50%

WITH-CARD VERSION PREFERRED

70with cards better
÷
107comparable tests
=
65%

The percentages use different groups: 50% counts continuations with text reporting card use among all 280 tests; 65% covers the 107 such cases with text on both sides. The generating model reports card use, while an anonymized AI judge compares writing quality.

Card use is self-reported by the generator. The anonymized AI judge sees the draft and source snippets, so source-aware preference remains possible; there was no independent human review. The overall result describes these 280 frozen positions, not a win probability for any writing task.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

Non-singular cosmological scenarios in scalar-tensor theories and their stability: a review

1 Introduction > 1.2 Horndeski theories as a suitable NEC-violating framework

Scalar-tensor theories embody probably the simplest way to extend GR, which amounts to introducing an additional scalar degree of freedom on top of the two tensor modes already present within the original GR. Even though the UV behaviour of the scalar-tensor theories is not much better than that of GR, the scalar-tensor family is of particular significance since most IR modifications of gravity can be described by scalar-tensor theories.

…n be cast into the form ( 5 ). In particular, upon a specific choice of scalar potentials F F , K K , G 4 G_{4} and G 5 G_{5} one can reproduce, for example, Einstein-Hilbert action, Brans-Dicke theory [ 29 ] , f ⁡ ( R ) f(R) -gravity theories [ 30 , 31 ] , k-essence [ 32 , 33 ] , kinetic gravity braiding [ 35 , 36 ] , Fab Four [ 37 , 38 ] (which involve couplings to the Gauss-Bonnet term), Dirac-Born-Infeld (DBI) galileons [ 39 , 40 ] , any inflationary theory based on a scalar field, e.g. k-inflation [ 41 ] and G-inflation [ 42 ] , and many more (see e.g. [ 44 , 43 ] for a thorough review).

↓ 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 positionOverall AI review: With cards better

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.

Read overall AI review rationale

A = without cards · B = with cards

The draft ends a long list of examples that can be cast into Horndeski form, with the last items being k-inflation and G-inflation. The natural next move is to comment on one of these examples or provide a thematic closing to this enumeration. Output B does this well: it picks up on the Gauss-Bonnet coupling mentioned in the Fab Four and DBI galileons entries, and notes its special status as the unique higher-curvature term with second-order equations in 4D—a well-known fact that connects to the generalized Galileon framework. This fits the current argument about the unifying power of Horndeski theory and flows naturally from the list just given. Output A, by contrast, abruptly pivots to non-singular cosmology and NEC violation, which is a completely new topic not prepared by the draft. The claim that Horndeski theories are 'a natural arena for constructing non-singular cosmological scenarios' is unsupported by the visible draft and constitutes a material topic drift. The draft is discussing the *scope* of Horndeski theory (what it can reproduce), not its applications to singularity resolution. Output A's claim about 'NEC-violating dynamics required to avoid the initial singularity' is a concrete unsupported assertion that is material to usability—it introduces a specific physical mechanism and cosmological application with no foundation in the preceding text.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

Among these, the Gauss-Bonnet coupling occupies a distinguished place: it is the unique higher-curvature term that yields second-order equations of motion in four dimensions, and it can be incorporated into the generalized Galileon framework in a non-trivial manner.

Passages supplied to the Evidence version

Card 1 · k-Inflation-corrected Einstein-Gauss-Bonnet gravity with massless primordial gravitons

In the present paper, we study the inflationary phenomenology of a k-inflation corrected Einstein-Gauss-Bonnet theory. Non-canonical kinetic terms are known for producing Jean instabilities or superluminal sound wave velocities in the aforementioned era, but we demonstrate in this work that by addin…

Read full passage excerpt

In the present paper, we study the inflationary phenomenology of a k-inflation corrected Einstein-Gauss-Bonnet theory. Non-canonical kinetic terms are known for producing Jean instabilities or superluminal sound wave velocities in the aforementioned era, but we demonstrate in this work that by adding Gauss-Bonnet string corrections and assuming that the non-canonical kinetic term ωXγ is in quadratic, one can obtain a ghost free description. Demanding compatibility with the recent GW170817 event forces one to accept that the relation ξ¨=Hξ˙ for the scalar coupling function ξ(ϕ). As a result, the scalar functions of the theory are revealed to be interconnected and by assuming a specific form for one of them, specifies immediately the other. Here, we shall assume that the scalar potential is directly derivable from the equations of motion, once the Gauss-Bonnet coupling is appropriately chosen, but obviously the opposite is feasible as well. As a result, each term entering the equations o …

Card 2 · Generalized G-Inflation: --Inflation with the Most General Second-Order Field Equations--

We study generalized Galileons as a framework to develop the most general single-field inflation models ever, Generalized G-inflation, containing yet further generalization of G-inflation, as well as previous examples such as k-inflation, extended inflation, and new Higgs inflation as special cases.…

Read full passage excerpt

We study generalized Galileons as a framework to develop the most general single-field inflation models ever, Generalized G-inflation, containing yet further generalization of G-inflation, as well as previous examples such as k-inflation, extended inflation, and new Higgs inflation as special cases. We investigate the background and perturbation evolution in this model, calculating the most general quadratic actions for tensor and scalar cosmological perturbations to give the stability criteria and the power spectra of primordial fluctuations. It is pointed out in the Appendix that the Horndeski theory and the generalized Galileons are equivalent. In particular, even the non-minimal coupling to the Gauss-Bonnet term is included in the generalized Galileons in a non-trivial manner.

Card 3 · Spherical cows in the sky with fab four

We explore spherically symmetric static solutions in a subclass of unitary scalar-tensor theories of gravity, called the `Fab Four' models. The weak field large distance solutions may be phenomenologically viable, but only if the Gauss-Bonnet term is negligible. Only in this limit will the Vainshtei…

Read full passage excerpt

We explore spherically symmetric static solutions in a subclass of unitary scalar-tensor theories of gravity, called the `Fab Four' models. The weak field large distance solutions may be phenomenologically viable, but only if the Gauss-Bonnet term is negligible. Only in this limit will the Vainshtein mechanism work consistently. Further, classical constraints and unitarity bounds constrain the models quite tightly. Nevertheless, in the limits where the range of individual terms at large scales is respectively Kinetic Braiding, Horndeski, and Gauss-Bonnet, the horizon scale effects may occur while the theory satisfies Solar system constraints and, marginally, unitarity bounds. On the other hand, to bring the cutoff down to below a millimeter constrains all the couplings scales such that `Fab Fours' can't be heard outside of the Solar system.

Card 4 · Constraining modified gravity with large non-Gaussianities

In writing a covariant effective action for single-field inflation, one is allowed to add a Gauss-Bonnet and axion-type curvature couplings. These couplings represent modifications of gravity, and are the unique higher-curvature terms that lead to second order equations of motion in four dimensions.…

Read full passage excerpt

In writing a covariant effective action for single-field inflation, one is allowed to add a Gauss-Bonnet and axion-type curvature couplings. These couplings represent modifications of gravity, and are the unique higher-curvature terms that lead to second order equations of motion in four dimensions. In this paper we study the observational consequences of such couplings for models with large non-Gaussianities. Our focus is on the Gauss-Bonnet term. In particular, we study an effective action where the scalar Lagrangian is a general function of the inflaton and its first derivative. We show that, for large non-Gaussianities, one can write ${f}_{\mathrm{NL}}$ in terms of only three parameters. The shape of ${f}_{\mathrm{NL}}$ is also studied, and we find that it is very similar to that of $k$-inflation. We show that the Gauss-Bonnet term enhances the production of gravitational waves, and allows a smaller speed of sound for scalar perturbations. This, in turn, can lead to larger non-Gaus …

Card 5 · Primordial perturbations and inflation in a holography inspired Gauss-Bonnet cosmology

We consider an action for gravity that, in addition to the Einstein-Hilbert term, contains a function of the Ricci scalar and the Gauss-Bonnet invariant. The specific form of the function considered is motivated by holographic cosmology. At background level the field equations imply modified Friedma…

Read full passage excerpt

We consider an action for gravity that, in addition to the Einstein-Hilbert term, contains a function of the Ricci scalar and the Gauss-Bonnet invariant. The specific form of the function considered is motivated by holographic cosmology. At background level the field equations imply modified Friedmann equations of the same form as those in the holographic cosmology. We calculate the cosmological perturbations and derive the corresponding power spectra assuming a general $k$-inflation. We find that the resulting power spectra differ substantially from those obtained in both holographic and standard cosmology. The estimated spectral index and tensor-to-scalar ratio are confronted with the Planck results.

Model-reported card use

Card 4 · fact

Affected output: it is the unique higher-curvature term that yields second-order equations of motion in four dimensions

Model explanation: This row supplied the specific property that singles out the Gauss-Bonnet coupling, which the draft's enumeration of Horndeski subclasses had not yet characterized.

Card 2 · fact

Affected output: it can be incorporated into the generalized Galileon framework in a non-trivial manner

Model explanation: This row supplied the fact that the Gauss-Bonnet coupling enters the generalized Galileon framework non-trivially, which the draft had not stated.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

This makes Horndeski theories a natural arena for constructing non-singular cosmological scenarios, since the additional scalar degree of freedom can support the NEC-violating dynamics required to avoid the initial singularity.

Earlier card-use comparison

The earlier diagnostic compared only cases with reported card use and text on both sides; its scope differs from the overall AI review above.

The blinded model review preferred the Evidence version. Its original rationale appears below.

Read the original model judgment

Output A introduces two unsupported claims: (1) that Horndeski theories are 'a natural arena for constructing non-singular cosmological scenarios' — none of the sources discuss non-singular cosmology or singularity avoidance; (2) that 'the additional scalar degree of freedom can support the NEC-violating dynamics required to avoid the initial singularity' — NEC violation and singularity avoidance are entirely absent from all supplied sources. The draft discusses scalar-tensor theories as unifying various gravity modifications, but never mentions cosmological singularities, NEC, or bounce scenarios. Output B is grounded: the claim that Gauss-Bonnet is 'the unique higher-curvature term that yields second-order equations of motion in four dimensions' is directly supported by the 'Constraining modified gravity with large non-Gaussianities' source ('the unique higher-curvature terms that lead to second order equations of motion in four dimensions'), and that it 'can be incorporated into the generalized Galileon framework in a non-trivial manner' is directly supported by the 'Generalized G-Inflation' source ('even the non-minimal coupling to the Gauss-Bonnet term is included in the generalized Galileons in a non-trivial manner'). Output B also naturally continues the draft's enumeration of specific theories by highlighting a distinguished structural property of one mentioned term.

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.