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 better45 / 280 · beu-0e3559782debe862

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

5 Conclusion

…2})}{X(3A_{1}X-4F_{2})}\;. (67) This is an exceptional subclass of DHOST Ia theories in which adding extra luminal scalar field does not necessarily leads to superluminality. Note that this subclass includes the theory with A 1 = 2 ​ F 2 ​ X A_{1}=2F_{2X} and A 3 = 0 A_{3}=0 , which is Horndeski, and does not include beyond Horndeski (GLPV) theories with X ​ A 3 = 2 ​ A 1 − 4 ​ F 2 ​ X XA_{3}=2A_{1}-4F_{2X} (the latter relation is inconsistent with ( 67 ) for 𝒢 𝒯 ≠ 0 \mathcal{G_{T}}\neq 0 , see eq. ( 47a )). This observation is in agreement with the result for beyond Horndeski in Sec. 4.3.1 .

Thus, the emergent superluminality over the cosmological background generally takes place in the DHOST Ia class upon adding to the system of even a tiny amount of kinetic energy Y > 0 Y>0 related to the additional scalar field χ \chi ( 55 ) with c m 2 = 1 c_{m}^{2}=1 . There is, however, an exceptional subclass of DHOST Ia theories given by ( 67 ), where the superluminality of perturbations in the scalar sector is not induced.

↓ 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 = with cards · B = without cards

Both outputs appropriately abstain with 'completion' actions, recognizing that the paragraph already makes a complete argumentative move. However, Output A is preferable because it precisely mirrors the source material's framing. The draft ends by identifying an exceptional subclass where superluminality is not induced; Output A directly echoes the source's explicit caveat about non-cosmological background consistency, which is the natural and exact next concern raised in the literature. Output B introduces broader, more speculative concerns ('embedding in a consistent ultraviolet completion') that go beyond what the sources mention and what the draft's argument structure calls for. The draft's scope is specifically about superluminality induced by adding kinetic energy for an extra scalar; Output A stays tightly coupled to this by noting the open question about background consistency, while Output B drifts toward more general theoretical concerns. Output A fits as a more precise, scope-appropriate closing remark that could be inserted without repair.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

The exceptional subclass (67) avoids the induced superluminality, but whether it remains consistent in non-cosmological backgrounds is still an open question.

Passages supplied to the Evidence version

Card 1 · Superluminality in DHOST theory with extra scalar

If one insists on the absence of superluminality, one has two logical possibilities. One of them is that in scalar-tensor theories with multiple scalars, none of the scalar fields has conventional kinetic term as long as there is at least one field of DHOST Ia type. Another is to stick to the except…

Read full passage excerpt

If one insists on the absence of superluminality, one has two logical possibilities. One of them is that in scalar-tensor theories with multiple scalars, none of the scalar fields has conventional kinetic term as long as there is at least one field of DHOST Ia type. Another is to stick to the exceptional subclass ( 19 ) of DHOST Ia theories and allow for conventional scalars. It remains to be understood, however, whether the second class of theories is consistent in non-cosmological backgrounds.

Card 2 · Cosmological evolution in DHOST theories

In the context of Degenerate Higher-Order Scalar-Tensor (DHOST) theories, we study cosmological solutions and their stability properties. In particular, we explicitly illustrate the crucial role of degeneracy by showing how the higher order homogeneous equations in the physical frame (where matter i…

Read full passage excerpt

In the context of Degenerate Higher-Order Scalar-Tensor (DHOST) theories, we study cosmological solutions and their stability properties. In particular, we explicitly illustrate the crucial role of degeneracy by showing how the higher order homogeneous equations in the physical frame (where matter is minimally coupled) can be recast in a system of equations that do not involve higher order derivatives. We study the fixed points of the dynamics, finding the conditions for having a de Sitter attractor at late times. Then we consider the coupling to matter field (described for convenience by a k-essence Lagrangian) and find the conditions to avoid gradient and ghost instabilities at linear order in cosmological perturbations, extending previous work. Finally, we apply these results to a simple subclass of DHOST theories, showing that de Sitter attractor conditions, no ghost and no gradient instabilities conditions (both in the self-accelerating era and in the matter dominated era) can be …

Card 3 · Stealth dark energy in scordatura DHOST theory

Abstract A stealth de Sitter solution in scalar-tensor theories has an exact de Sitter background metric and a nontrivial scalar field profile. Recently, in the context of Degenerate Higher-Order Scalar-Tensor (DHOST) theories it was shown that stealth de Sitter solutions suffer from either infinite…

Read full passage excerpt

Abstract A stealth de Sitter solution in scalar-tensor theories has an exact de Sitter background metric and a nontrivial scalar field profile. Recently, in the context of Degenerate Higher-Order Scalar-Tensor (DHOST) theories it was shown that stealth de Sitter solutions suffer from either infinite strong coupling or gradient instability for scalar field perturbations. The sound speed squared is either vanishing or negative. In the first case, the strong coupling scale is zero and thus lower than the energy scale of any physical phenomena. From the viewpoint of effective field theory, this issue is naturally resolved by introducing a controlled detuning of the degeneracy condition dubbed scordatura, recovering a version of ghost condensation. In this paper we construct a viable dark energy model in the scordatura DHOST theory based on a stealth cosmological solution, in which the metric is the same as in the standard ΛCDM model and the scalar field profile is linearly time-dependent. …

Card 4 · CMB constraints on DHOST theories

Abstract We put constraints on the degenerate higher-order scalar-tensor (DHOST) theories using the Planck 2018 likelihoods. In our previous paper, we developed a Boltzmann solver incorporating the effective field theory parameterised by the six time-dependent functions, αi (i = B, K, T, M, H) and β…

Read full passage excerpt

Abstract We put constraints on the degenerate higher-order scalar-tensor (DHOST) theories using the Planck 2018 likelihoods. In our previous paper, we developed a Boltzmann solver incorporating the effective field theory parameterised by the six time-dependent functions, αi (i = B, K, T, M, H) and β 1, which can describe the DHOST theories. Using the Markov-Chain Monte-Carlo method with our Boltzmann solver, we find the viable parameter region of the model parameters characterising the DHOST theories and the other standard cosmological parameters. First, we consider a simple model with α = ΩDE(t)/ΩDE(t 0), α B = α T = α M = α H = 0 and β 1 = β 1,0ΩDE(t)/ΩDE(t 0) in the ΛCDM background where t 0 is the present time and obtain β 1,0 = 0.032-0.016 +0.013 (68% c.l.). Next, we focus on another theory given by ℒDHOST = X + c 3 X□ϕ/Λ3 + (M pl 2/2 + c 4 x 2/Λ6)R + 48c 4 2 X 2/ (M pl 2Λ12 + 2c 4Λ6 X 2)ϕμ ϕμρ ϕρν ϕν with X := ∂μϕ∂μϕ and two positive constant parameters, c 3 and c 4. In this mode …

Card 5 · Superluminality in DHOST theory with extra scalar

If one insists on the absence of superluminality, one has two logical possibilities. One of them is that in scalar-tensor theories with multiple scalars, none of the scalar fields has conventional kinetic term as long as there is at least one field of DHOST Ia type. Another is to stick to the except…

Read full passage excerpt

If one insists on the absence of superluminality, one has two logical possibilities. One of them is that in scalar-tensor theories with multiple scalars, none of the scalar fields has conventional kinetic term as long as there is at least one field of DHOST Ia type. Another is to stick to the exceptional subclass ( 19 ) of DHOST Ia theories and allow for conventional scalars. It remains to be understood, however, whether the second class of theories is consistent in non-cosmological backgrounds.

Model-reported card use

Card 1 · fact

Affected output: whether it remains consistent in non-cosmological backgrounds is still an open question

Model explanation: The row supplies the specific open question about consistency of the exceptional subclass in non-cosmological backgrounds, which the draft's conclusion did not yet state.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

The stability of these exceptional configurations under perturbations beyond the cosmological background, and their embedding in a consistent ultraviolet completion, remain open questions.

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 is grounded and directly supported by the source material. The source explicitly states about the exceptional subclass: 'It remains to be understood, however, whether the second class of theories is consistent in non-cosmological backgrounds.' Output A faithfully paraphrases this exact point, maintaining the scope (non-cosmological backgrounds) and the epistemic status (open question). Output B introduces two unsupported claims: (1) 'stability of these exceptional configurations under perturbations beyond the cosmological background' — the source mentions consistency in non-cosmological backgrounds, not stability under perturbations beyond cosmological backgrounds; and (2) 'their embedding in a consistent ultraviolet completion' — this is entirely absent from all supplied sources. The UV completion claim is a material factual addition with no source support, making Output B contain unsupported claims and therefore unusable.

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.