FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteEvidence version preferred64 / 162 · b9d645aa630589a5

Survey of Consciousness Theory from Computational Perspective

Psychology and cognitive neuroscience · 2309.10063v1

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

Survey of Consciousness Theory from Computational Perspective

1 Introduction > 1.2 Definition of Consciousness

The above conversation is a microcosm of the discussions by the authors of the paper for starting the investigation of the consciousness problem. We will start the discussion with the definition of consciousness (Sec. 1.2 ), and the relationship between consciousness and intelligence (Sec. 1.4 ), consciousness and free will (Sec. 1.5 ). Then we have a brief overview of the existing consciousness theories (Sec. 1.7 ) and a further discussion about consciousness while asleep (Sec. 1.6 ).

Consciousness, Awareness, Wakefulness and Arousal: Consciousness is a complex and multifaceted concept that has been studied extensively in the fields of neuroscience, psychology, and philosophy. Existing research typically defines consciousness as comprising two main components: arousal (wakefulness) and awareness (subjective experience) ( Lendner et al., 2020 ) . Arousal refers to the overall state of alertness or wakefulness, while awareness refers to the subjective experience of perceiving and interpreting sensory information.

↓ 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: 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

Preferred by blind review

This two-dimensional framework has been used to classify disorders of consciousness, in which wakefulness and awareness can dissociate: patients in a vegetative state show wakefulness without awareness, whereas patients in a minimally conscious state retain intermittent but limited awareness.

Passages supplied to the Evidence version

The Morphospace of Consciousness ↗

This clinical definition of consciousness enables a practical classification of closely associated states/disorders of consciousness into clusters on a bivariate scale with awareness and wakefulness on orthogonal axes. Under healthy conditions, these two levels are almost linearly correlated, as in…

Read full passage excerpt

This clinical definition of consciousness enables a practical classification of closely associated states/disorders of consciousness into clusters on a bivariate scale with awareness and wakefulness on orthogonal axes. Under healthy conditions, these two levels are almost linearly correlated, as in conscious wakefulness (high arousal and high awareness) or in deep sleep (low arousal and low awareness). In pathological states, wakefulness without awareness can be observed in the vegetative state [ 1 ] , while transiently reduced awareness is observed following seizures [ 18 ] . Patients in the minimally conscious state show intermittent and limited non-reflexive and purposeful behavior [ 19 ] , [ 20 ] , whereas patients with hemispatial neglect display reduced awareness of stimuli contralateral to the side where brain damage has occurred [ 21 ] .

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

This distinction is central to clinical and theoretical accounts, where disorders of consciousness are often characterized by dissociations between the two: patients may retain wakefulness without awareness, as in the vegetative state, or show awareness without corresponding behavioral responsiveness, as in covert cognition.

Judgment and task check

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

Read the original model judgment

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

Both outputs provide coherent, source-grounded continuations. Output B more precisely matches the source: it specifies 'two-dimensional framework' (aligning with 'bivariate scale with awareness and wakefulness on orthogonal axes'), correctly identifies 'vegetative state' as 'wakefulness without awareness,' and 'minimally conscious state' as having 'intermittent but limited awareness'—both directly from the source snippet. Output A introduces 'covert cognition' and 'awareness without corresponding behavioral responsiveness,' which are not mentioned in the supplied source. While not necessarily false, 'covert cognition' is an unsupported extrapolation in this closed-book evaluation. Output B stays strictly within the source's scope and terminology, making it the more defensible choice for direct insertion with less risk of unsupported claims.

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.