FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteJudgment incomplete92 / 162 · 7afb345de64d3c79

Predictive Coding: a Theoretical and Experimental Review

Psychology and cognitive neuroscience · 2107.12979v4

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

Predictive Coding: a Theoretical and Experimental Review

1 Introduction

… Seth, (2014) ) . Cybernetics as a science is focused on understanding the dynamics of interacting feedback loops for perception and control, based especially around the concept of error minimization. Control and filtering theory have, in a related but distinct way, been based around methods to minimize residual errors in both perception or action according to some objective for decades. As we shall see, standard methods such as Kalman Filtering ( Kalman, (1960) ) or PID control Johnson & Moradi ((2005)) can be shown as special cases of predictive coding under certain restrictive assumptions.

…zes the redundancy in the representation of the visual information closely resembles the centre-surround receptive fields which are well established in retinal ganglion cells ( Huang & Rao, (2011) ) . This predictive coding approach was also applied to coding in the lateral geniculate nucleus (LGN), the thalamic structure that retinal signals pass through en-route to cortex, which was hypothesised to help remove temporal correlations in the input by subtracting out the retinal signal at previous timesteps using recurrent lateral inhibitory connectivity ( Huang & Rao, (2011) ; Marino, (2020) )

↓ 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: Judgment incomplete

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

Generation status: failed. No complete output is available for this arm.

Passages supplied to the Evidence version

Meaningful representations emerge from Sparse Deep Predictive Coding ↗

Sparse coding strategies have been successfully applied by neuroscientists to model some properties of the visual cortex [ 9 , 10 ] . Recent advances have shown that these models are related to a wide range of mathematical methods to extract sparse representations [ 11 ] , yet most often with a sing…

Read full passage excerpt

Sparse coding strategies have been successfully applied by neuroscientists to model some properties of the visual cortex [ 9 , 10 ] . Recent advances have shown that these models are related to a wide range of mathematical methods to extract sparse representations [ 11 ] , yet most often with a single processing layer. Interestingly, Predictive Coding (PC), a theory formalized by Rao & Ballard [ 12 ] to model contextual effects in a population of V1 neurons, gives a framework to combine sparse coding with a hierarchical structure [ 13 ] . Indeed, PC suggests that the brain uses Bayesian inference to consistently update an internal model of the world that describes the possible causes of a given sensory input [ 14 ] . Whereas the vast majority of CNN are mainly feed-forward, PC introduces recurrent connections combining top-down and bottom-up processes.

No Evidence

Same model and position · no retrieved passages · first output

Predictive coding thus sits at the intersection of these traditions, offering a normative account in which perception and action both follow from the minimization of prediction error.

Judgment and task check

The judgment was incomplete, so this case has no winner but remains in the denominator above.

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.