Chris von Csefalvay is a computational epidemiologist/data scientist working at the intersection of AI, epidemiology and public health.

Chris von Csefalvay

It appears that in what is clearly a wonderful little PR stunt, a Polish rum company managed to do a Sophia and appoint an ‘AI-driven’ ‘robot’ as its ‘CEO’.  The other guilty party to this pile of steaming bovine excrement is Hanson Robotics, famous for giving us Sophia, the “world’s first robot citizen”. Most of what I’m saying here goes just as well for Sophia.

The skeuomorphic fallacy

Over at the work blog, I’m discussing what knowledge means for large language models (LLMs), and the ways in which we can leverage this knowledge dividend for better inference.  Read the full post here.

Asemantic Induction of Hallucinations in LLMs

There’s a style of visual design I’m inordinately fond of called Raygun Gothic. It’s hard to describe what the hell exactly one needs to be on to enjoy it, but think of it like the aesthetic from the latter Fallout games with a more optimistic outlook on the future. Gibson described it as “the future that never was”, and I think that’s a pretty apt description. We get these competing futures in technology all the time.

Prompt Engineering: The Art of Yesterday

Say you’re busing tables and you’re trying to pass someone in a wheelchair. What do you do? Do you say “excuse me” and wait for them to move? Do you say “excuse me” and then try to pass them? Do you just try to pass them? Do you say nothing and just try to pass them? All of these are, actually, pretty legitimate answers. Now, say you’re a robot.

What I learned from getting bodied by a robot.

There’s a notion in artificial intelligence known as Moravec’s paradox: it’s relatively easy to teach a computer to play chess or checkers at a pretty decent level, but near impossible to teach it something as trivial as bipedal motion. (Hassabis 2017) The sensorimotor tasks that our truly wonderful brains have mastered by our second birthday are much harder to teach a computer than something arguably as ‘complex’ as beating a chess grandmaster.

The hardest AI problem you’ve never heard of.

The year is 1959. Eisenhower is on his second term, Castro just kicked Batista out of the country and Ray Charles’s Let the Good Times Roll is topping the charts.

The 95% myth

There’s something special about language. It is ‘our own’, it is ‘us’, in a profound way, and quite surprisingly, more so than art. I was deeply struck by this when I first saw reactions to large generative language models that created realistic, human-ish prose.

Beyond Broca

The knowledge dividend of large language models

In my latest post for the Starschema blog, I discuss the end of dashboards, and what comes next:  Read the full post here.

As we may see: the world after dashboards

There might be valid concerns about the mRNA vaccines. Then there are the clearly insane ones, such as the claims that it involves the injection of a microchip. The argument that mRNA vaccines are a form of ‘gene therapy’ are somewhere in the middle, and that makes responding to it somewhat difficult.

Why the COVID-19 vaccines are not gene therapy

The awesome thing about language is that, well, we all mostly speak it, to some extent or another. This gives us an immensely powerful tool to manipulate transformational tasks. For the purposes of this post, I consider a transformational task to be essentially anything that takes an input and is largel intended to return some version of the same thing. This is not a very precise definition, but it will have to do for now.

LAIR - Language As Intermediate Representation

In the first four entries (1 2 3 4) of this sequence, I have focused primarily on what LLMs aren’t, can’t, won’t, wouldn’t and shouldn’t. It’s probably time to conclude this series by that much awaited moment in all stories, where the darkest night finally turns into a glorious dawn, where we finally arrive at the promised land, where we finally get to talk about what LLMs
<em>
 could
</em>
be.  What I see as the most successful potential model of

Teams of Rivals

Tip  If you’re here for the DOI matching script, the link is here. Quarto is a great tool for reproducible research. It is also a great tool for building websites. After about a decade of Wordpress-based websites, I’ve moved to Quarto primarily for two reasons. First, most of my content is largely static.

Quarto project scripts are awesomeness

It’s not every day that you find out you have climbed the exalted heights of another discipline. My work is pretty interdisciplinary, but it shocked me, too, that I’m apparently holding forth on neoliberalism and the epistemic question in African universities (archive link):  This, of course, came at some surprise to me, as I have never written anything on the topic.

Stochastic parrots, cap and gown edition

The temple of Asclepios, the Greek god of healing arts and medicine, at Epidaurus was pretty much the ancient Greek world’s equivalent of the Mayo Clinic. It then tells us a lot about the Greek worldview of healing that one of the things the temple complex prominently featured was a theatre. The Greeks believed in drama therapy, but in perhaps a slightly different way of how we think of it today.

The Moral Pulse of the Machine

I have probably spent more time looking at Poussin’s
<em>
 Dance to the Music of Time
</em>
than any other work of art. Sneaking off to the Wallace Collection in London and just looking at the
<em>
 Dance
</em>
was my comfort activity while living in London – a time that was not exactly devoid of its trials. It’s not, by any measure,
<em>
 great
</em>
art, insofar as such judgments can be made with any objectivity.

The Lyre of Hephaestus

My recent post on the Starschema blog discusses the need for better data products to tackle future pandemic challenges:  Read the full post here.

Data for the next pandemic

In a recent paper that has attracted the interest of popular media as well, Fabio Urbina and colleagues examined the use (or rather, the abuse) of computational chemistry models of toxicity for generating toxic compounds and potential chemical agent candidates.(Urbina et al. 2022) Urbina and colleagues conclude that Urbina, Fabio, Filippa Lentzos, Cédric Invernizzi, and Sean Ekins. 2022.

A different shade of grey

Every year, the Commandant of the Marine Corps publishes a reading list of books that often only bear on warfighting tangentially at best. The idea behind this is that those entrusted with the lives of servicemembers should have an understanding of the world that goes beyond the profession of arms. In much the same way, I have been advising data scientists to go beyond professional literature.

Five non-data-science books for data scientists

I used to be on Quora. I’m not anymore. But I used to be. And I wrote some answers. Here are some of my favourites, categorised by field… sort of.
<strong>
 AI
</strong>
If I wrote an AI and trained it to score 200 on a standardized IQ test, what could a human of 100 IQ still do better?   Read the full answer here. Will true AI come before the ability to completely simulate a human brain? Is there a difference?   Read the full answer here.

My favourite Quora answers

As the new Omicron variant of concern of COVID-19 emerged over Thanksgiving week (of all times!), attention has been turning to Paxlovid (PF-07321332). If, like me, you torment yourself by perusing conspiracy theories, you might have heard it referred to as Pfizermectin.

Peace, love, Paxlovid and ‘Pfizermectin’

In the middle of the green and pleasant countryside of England lies a small piece of commons –– land, that is, over which a number of people share certain rights. In this hypothetical scenario, it’s six farmers,  to . Each of them has a hypothetical cow, which differs from ordinary cows in that it grazes a fixed unit, and produces a fixed amount of milk (unlike real cows, which vary wildly in their input and output parameters, or so I’m told).

Ransomware attacks as n-person prisoner’s dilemmas

Yesterday, I completed a marathon on the SkiErg, with a pretty respectable time (see results here). I know I can do better, and I know I
<em>
 will
</em>
do better. On the other hand, this was a marathon where I pretty much did everything wrong. And that was sort of an instructive experience in its own right. The title is somewhat misleading. Not much went wrong – the marathon itself was smooth sailing.

How to do a SkiErg marathon entirely the wrong way (but still finish)

I love posts that allow me to merge some of my addictions.

Auto-DOI for Quarto posts via Rogue Scholar

The awesome thing about language is that, well, we all mostly speak it,
to some extent or another. This gives us an immensely powerful tool to
manipulate transformational tasks. For the purposes of this post, I
consider a transformational task to be essentially anything that takes
an input and is largel intended to return some version of the same
thing. This is not a very precise definition, but it will have to do for
now.

Such models are nothing new. Perhaps the most eye-catching and
ubiquitous of such models are [neural style
transfer](https://www.tensorflow.org/tutorials/generative/style_transfer)
models that take an image and return a version of the same image in a
different style that let you turn a picture of your dog into a Van Gogh
painting [(Gatys, Ecker, and Bethge 2015)]{.citation
cites="gatys2015neural"}. Quite simply put, these models are your
typical generative model, with the difference that it takes two separate
loss definitions: content loss, which is loss of the generated image
vis-a-vis the content reference, and style loss, which is the loss
vis-a-vis the style reference image. A "good" image then is one that
minimises total loss, i.e. it's just as close to your dog as it is to
Van Gogh. Figure 1 outlines this logic.

::: {.no-row-height .column-margin .column-container}
::: {#ref-gatys2015neural .csl-entry}
Gatys, Leon A, Alexander S Ecker, and Matthias Bethge. 2015. 'A Neural
Algorithm of Artistic Style'. *arXiv Preprint arXiv:1508.06576*.
:::
:::

::: {.cell layout-align="center"}
::: {.cell-output-display}
::: {#fig-nst .quarto-figure .quarto-figure-center .anchored}
<figure class="figure">
<div>
<pre class="mermaid mermaid-js" data-label="fig-nst"><code>flowchart TD
    C[&quot;Content image&quot;]
    S[&quot;Style image&quot;]

    G[&quot;Generated image&quot;]

    G --&gt; L[&quot;Loss network&quot;]
    L --&gt; G
    
    C -- &quot;Content loss&quot; --&gt; L
    S -- &quot;Style loss&quot; --&gt; L</code></pre>
</div>
<figcaption>Figure 1: A rough outline of NST.</figcaption>
</figure>
:::
:::
:::

Given a content source image
![](https://latex.codecogs.com/png.latex?%5Cvec%7Bc%7D) and a style
reference image ![](https://latex.codecogs.com/png.latex?%5Cvec%7Bs%7D),
we define the total loss of our generated image
![](https://latex.codecogs.com/png.latex?%5Cvec%7Bg%7D) as

![](https://latex.codecogs.com/png.latex?%0A%20%20%20%20%5Cmathcal%7BL%7D_%7Btotal%7D%20=%20%5Calpha%20%5Cmathcal%7BL%7D_%7Bcontent%7D%20+%20%5Cbeta%20%5Cmathcal%7BL%7D_%7Bstyle%7D%0A)

where ![](https://latex.codecogs.com/png.latex?%5Calpha) and
![](https://latex.codecogs.com/png.latex?%5Cbeta) are hyperparameters
that control the relative importance of the content and style losses.
The content loss for layer ![](https://latex.codecogs.com/png.latex?l)
is defined as

![](https://latex.codecogs.com/png.latex?%0A%20%20%20%20%5Cmathcal%7BL%7D_%7Bcontent%7D(%5Cvec%7Bc%7D,%20%5Cvec%7Bg%7D,%20l)%20=%20%5Cfrac%7B1%7D%7B2%7D%20%5Csum_%7Bi,j%7D%20(%5Cvec%7Bc%7D_%7Bij%7D%5El%20-%20%5Cvec%7Bg%7D_%7Bij%7D%5El)%5E2%0A)

which is basically a simple squared error loss between the feature
vector of the content image and the generated image at layer
![](https://latex.codecogs.com/png.latex?l). The style loss is a bit
more complicated, and is these days typically defined as the Maximum
Mean Discrepancy, which [Li et al. (2017)]{.citation
cites="li2017demystifying"} have shown is essentially equivalent to the
Gram matrix loss, defined as

::: {.no-row-height .column-margin .column-container}
::: {#ref-li2017demystifying .csl-entry}
Li, Yanghao, Naiyan Wang, Jiaying Liu, and Xiaodi Hou. 2017.
'Demystifying Neural Style Transfer'.
<https://arxiv.org/abs/1701.01036>.
:::
:::

![](https://latex.codecogs.com/png.latex?%0A%20%20%20%20%5Cmathcal%7BL%7D_%7Bstyle%7D(%5Cvec%7Bs%7D,%20%5Cvec%7Bg%7D,%20l)%20=%20%5Cfrac%7B1%7D%7B4N_l%5E2M_l%5E2%7D%20%5Csum_%7Bi,j%7D%20(%5Cmathbf%7BG%7D_%7Bij%7D%5El%20-%20%5Cmathbf%7BS%7D_%7Bij%7D%5El)%5E2%0A)

where
![](https://latex.codecogs.com/png.latex?%5Cmathbf%7BG%7D_%7Bij%7D%5El)
and
![](https://latex.codecogs.com/png.latex?%5Cmathbf%7BS%7D_%7Bij%7D%5El)
are the Gram matrices of the generated reference image and the style
image at layer ![](https://latex.codecogs.com/png.latex?l),
respectively. ![](https://latex.codecogs.com/png.latex?N%5El) is the
number of feature maps in layer
![](https://latex.codecogs.com/png.latex?l) and
![](https://latex.codecogs.com/png.latex?M_l) is the dimensionality
(height times width) of the feature map of layer
![](https://latex.codecogs.com/png.latex?l). There are two fundamental
problems with this.

1.  This works much less well for things that aren't images or at least
    sufficiently similar to images.
2.  The bigger problem is that the style reference is pretty much
    exhaustive. By that, I mean that there isn't much we can convey to
    the model about the style that isn't encapsulated in the style
    reference images. Depending on how semantically apt your model is,
    it may or may not be able to pick up some higher level ideas. It may
    be able to pick up the brush strokes of Van Gogh or the colours of a
    Turner, but it may not be able to paint your characters in the
    semantic context of Van Gogh's time and place.

This is where language comes in. Language is a very powerful tool for
conveying information, and it turns out that if we use language as an
intermediate representation, we can use language models to manipulate
this intermediate representation to our heart's content, using a
relatively informal and rather forgiving language. This is the idea
behind LAIR.

::: {#a-toy-example .section .level2 .page-columns .page-full}
## A toy example {.anchored anchor-id="a-toy-example"}

::: {.page-columns .page-full}
I'm a firm believer in silly toy examples. The sillier, the better. So,
we'll start with the following proposition: can we create a model that
will look at a photo from, say, the front page of our favourite
newspaper, and transpose it into the Warhammer 40k universe? In case
you're unfamiliar, Warhammer 40k is set -- as the name suggests -- in
the 40th millennium, but is a weird mixture of medieval and futuristic
technology. The whole atmosphere is taking the 'Dark Ages' part of the
Middle Ages,^1^ adding spaceships and laser guns, and turning the whole
thing into an absolutely depressing dystopia. It's a lot of fun.

::: {.no-row-height .column-margin .column-container}
^1^ Which may or may not ever have actually existed.
:::
:::

There are a few things we want here: I don't merely want the *visual*
style of the Warhammer 40k universe, I also want the *semantics* -- that
is, I want characters to be transposed into the Warhammer 40k universe.
I want the model to understand that the people in the photo are now
Space Marines, and that the buildings are now Gothic cathedrals. I can't
get neural transfer to that for me, because it does not understand, or
care, about semantics, and does not do semantic transformation. More
importantly, I cannot interact with the 'guts' of neural style transfer
beyond setting the hyperparameters and the source images.

What I can, however, do is to use the language I am mostly most
proficient in -- that is, human language -- to manipulate an
intermediate representation.

::: {.cell layout-align="center"}
::: {.cell-output-display}
::: {#fig-lair .quarto-figure .quarto-figure-center .anchored}
<figure class="figure">
<div>
<pre class="mermaid mermaid-js" data-label="fig-lair"><code>flowchart LR
    subgraph Description
        direction TB
        S[&quot;Source image&quot;] --&gt; D(&quot;Descriptor\ne.g. GPT-4 vision&quot;) --&gt; d[&quot;Description&quot;]
    end

    subgraph Transformation
        direction TB
        t(&quot;Transformer\ne.g. GPT-4&quot;) --&gt; td[&quot;Transformed\ndescription&quot;]
    end

    subgraph Rendering
        direction TB
        r(&quot;Renderer\ne.g. DALL-E&quot;) --&gt; I[&quot;Output\nimage&quot;]
    end

    Description --&gt; Transformation --&gt; Rendering</code></pre>
</div>
<figcaption>Figure 2: A rough outline of LAIR as applied to the toy
example.</figcaption>
</figure>
:::
:::
:::

Figure 2 outlines the basic idea for images. We use a descriptor model
to describe the image, and then use a text-to-text transformer to
transform the description. The transformed description is then rendered
into an image. The descriptor model is typically a vision model, and the
transformer is typically a language model. The renderer would then be a
generative model.

Consider the following source image I downloaded from Pexels:

::: {#fig-source .quarto-figure .quarto-figure-center .anchored}
<figure class="figure">
<p><img
src="https://chrisvoncsefalvay.com/posts/lair/pexels-canva-studio-3153198.jpg"
class="img-fluid figure-img" /></p>
<figcaption>Figure 3: Our source image.</figcaption>
</figure>
:::

Here's the intermediate representation (IR) description created by GPT-4
vision:

> This is an image of four individuals gathered around a laptop,
> suggesting a collaborative work environment. The setting has an ample
> amount of natural light, evident by the bright illumination on the
> subjects and the interior. They appear to be located in a modern
> indoor environment with a large curved window in the background,
> contributing to the light-flooded ambiance.
>
> Starting from the left, the first individual is a man with short, dark
> hair and a light beard, wearing a black t-shirt with a white graphic
> and text. He appears focused, resting his hands on the laptop while
> looking at the screen. To his right is a woman with shoulder-length
> curly hair who seems engaged and smiling, leaning slightly towards the
> laptop. She's wearing a light grey short-sleeved top.
>
> Behind these two and slightly out of focus, two more individuals are
> present. The person in the middle of the composition is a man with
> shoulder-length wavy hair, wearing a black shirt, gazing at the laptop
> screen. To the far right, partially cropped out of the frame, there is
> a woman with glasses, wearing a dark business casual blazer. She seems
> to be observing the laptop's screen with interest, standing slightly
> behind the seated individuals.
>
> The group exhibits a sense of teamwork and focus, with everyone's
> attention directed toward the laptop, situated on what appears to be a
> white table or desk. The body language and expressions suggest a
> positive, collaborative atmosphere, typical of a professional or
> academic team setting.

This is a pretty decent description. Let's grimdark it up by passing it
to our transformer model, which was initialised with the following
prompt:

> Your job is to analyze descriptions of images and change them to
> transpose the image into the dark, futuristic aesthetic of the
> Warhammer 40k universe. You are given a description of a scene, and
> your job is to change that scene according to the instructions, and
> give a detailed description of the changed scene that would allow a
> generative model to create an image.
>
> 1.  Replace clothing items with a mixture of futuristic and mediaeval
>     looking armour.
> 2.  Replace every-day objects held by characters with power swords,
>     chainswords,futuristic looking guns, staffs or magic items.
> 3.  Replace architecture with the monumental, dark architecture common
>     to the Warhammer 40k universe.
> 4.  Include things like monumental gothic spaceships in the image
>     description.

Here's the output:

> This is an image of four battle-hardened individuals gathered around a
> hololith, suggesting a strategic war council. The setting is bathed in
> the cold, harsh light of flickering lumens, casting long, grim shadows
> on the figures and the interior. They appear to be situated in a
> cathedral-like command room of an Imperial Battleship, with a
> monumental, gothic archway in the background, contributing to the
> oppressive ambiance.
>
> Starting from the left, the first individual is an Adeptus Astartes,
> his short, dark hair and light beard visible beneath the rim of his
> power armor helmet. He is clad in a black power armor with white
> Imperial insignia. He appears focused, his gauntleted hands resting on
> the hololith while studying the battlefield projection. To his right,
> a Sororitas, her shoulder-length, curly hair peeking out from beneath
> her wimple, leans towards the holographic display. She's clad in a
> light grey power armor, her face engaging and determined.
>
> Behind these two and slightly out of focus, two more figures loom. The
> figure in the middle of the composition is a man, his shoulder-length
> wavy hair barely visible beneath his Inquisitorial hat, wearing a
> black, ornate, medieval-looking armor, his eyes fixed on the hololith.
> To the far right, partially cropped out of the frame, there is a
> Tech-Priestess, her bionic glasses glinting in the lumen-light, her
> dark, mechanicus robes adding to her awe-inspiring presence. She seems
> to be observing the hololith with keen interest, standing slightly
> behind the armored individuals.
>
> The group gives off a sense of unity and focus, with everyone's
> attention directed toward the hololith, situated on what appears to be
> an aged, white marble table. The body language and expressions suggest
> a tense, collaborative atmosphere, typical of a war council in the
> grim darkness of the far future.

By the Emperor, that's some good stuff. Let's see what DALL-E makes of
it:

::: {#fig-output .quarto-figure .quarto-figure-center .anchored}
<figure class="figure">
<p><img
src="https://chrisvoncsefalvay.com/posts/lair/img-eTqdxb5CYGT2vY5FqNkRJdva.png"
class="img-fluid figure-img" /></p>
<figcaption>Figure 4: Our output image.</figcaption>
</figure>
:::

It's interesting to note what remains and what has changed. In
particular, the relationship of the figures to each other, both
spatially and semantically, as well as their posture, their number
(mostly!) and the general subject matter at a highly semanticised level
was preserved. Instead of a laptop, they're looking at a big holographic
sand table. Note that unlike in NST, we actually have control over what
is preserved and what is not -- that's what our transformer prompt
accomplishes. We've told it to change dress, for instance, but we
haven't told it to make any changes to the overall relationship between
the figures (note how a "collaborative atmosphere", for instance, was
retained word for word).
:::

::: {#whats-the-point .section .level2 .page-columns .page-full}
## What's the point? {.anchored anchor-id="whats-the-point"}

Okay, this was quite fun, but what's the point? Well, the point is that
we can use language as an intermediate representation to manipulate
things relatively easily. This is powerful because language is such an
accessible and forgiving intermediate representation. More importantly,
however, *anything* that can be represented in language can be
manipulated this way.

- In the space of language-to-language, this includes adapting text to
  the needs of special audiences,[Verma, Boland, and Miesenberger
  (2023)]{.citation cites="steinmetz2023developing"} conveying
  information to lay audiences [(Lee, Goldberg, and Kohane
  2023)]{.citation cites="lee2023ai"} and creating secondary explanatory
  materials e.g. statutory explanations [(Blair-Stanek, Holzenberger,
  and Van Durme 2023)]{.citation cites="blair2023can"}.
- For image-to-image transformations, retaining semanticity while
  simpifying visuals and removing clutter may often be useful, e.g. for
  creating procedural visual guidance for medical procedures [(Chen
  2023)]{.citation cites="chen2023endoscopic"}. Often, such images are
  created by hand, but this is a time-consuming process that could be
  automated.
- For code-to-code, language as an intermediate representation allows
  the interjection of desired features into code, e.g. for the purposes
  of code refactoring. Beyond simple code rewriting, this allows a kind
  of opinionated transformation. Often, a target language is not only
  idiomatically different but also has certain other characteristics,
  and this is *a fortiori* the case for DSLs [(Magalhães et al.
  2023)]{.citation cites="magalhaes2023c2taco"}.
- For code-to-text, this allows the creation of documentation from code,
  which is a perennial problem in software engineering. The textual
  intermediate representation allows fine control over the resulting
  documentation.

::: {.no-row-height .column-margin .column-container}
::: {#ref-verma2023bridging .csl-entry}
Verma, A Kumar, S Gavra Boland, and Klaus Miesenberger. 2023. 'Bridging
the Digital Divide for Persons with Intellectual Disabilities: Assessing
the Role of ChatGPT in Enabling Access, Evaluation, Integration,
Management, and Creation of Digital Content'. In *ICERI2023
Proceedings*, 3767--76. IATED.
:::

::: {#ref-lee2023ai .csl-entry}
Lee, Peter, Carey Goldberg, and Isaac Kohane. 2023. *The AI Revolution
in Medicine: GPT-4 and Beyond*. Pearson.
:::

::: {#ref-blair2023can .csl-entry}
Blair-Stanek, Andrew, Nils Holzenberger, and Benjamin Van Durme. 2023.
'Can GPT-3 Perform Statutory Reasoning?' *arXiv Preprint
arXiv:2302.06100*.
:::

::: {#ref-chen2023endoscopic .csl-entry}
Chen, Hao-Wen. 2023. 'Endoscopic Endonasal Skull Base Surgery for
Pituitary Lesions: An AI-Assisted Creative Workflow to Develop an
Animated Educational Resource for Patients and Physicians'. PhD thesis,
Johns Hopkins University.
:::

::: {#ref-magalhaes2023c2taco .csl-entry}
Magalhães, José Wesley de Souza, Jackson Woodruff, Elizabeth Polgreen,
and Michael FP O'Boyle. 2023. 'C2TACO: Lifting Tensor Code to TACO'. In
*Proceedings of the 22nd ACM SIGPLAN International Conference on
Generative Programming: Concepts and Experiences*, 42--56.
:::
:::

The possibilities for using language as an intermediate representation
are endless. LAIR is a powerful paradigm for transformational tasks that
allows us to use language as an intermediate representation to
manipulate things in a way that is both accessible and powerful, and
that allows us to pick and choose *what* part of semanticity we want to
manipulate *versus* what we want to preserve. It's hard to reason about
LAIR's relative performance given that it is not a technique but a
paradigm, and that its focus is not simple style transfer but finely
controlled stylistic and contextual transformation, but even in the
current absence of benchmarks, it is clear that models benefit from
using language as an easily workable and malleable intermediate
representation.
:::

::: {#code .section .level2}
## Code {.anchored anchor-id="code"}

The code for the toy example is available
[here](https://gist.github.com/chrisvoncsefalvay/f3c6e7547350cfae50be0078ad99435f).
:::

::: {#quarto-appendix .default}
::: {.section .quarto-appendix-contents}
## Citation {#citation .anchored .quarto-appendix-heading}

<div>

::: {.quarto-appendix-secondary-label}
BibTeX citation:
:::

``` {.sourceCode .code-with-copy .quarto-appendix-bibtex}
@online{csefalvay2024,
  author = {Chris von Csefalvay},
  title = {LAIR - {Language} {As} {Intermediate} {Representation}},
  date = {2024-01-06},
  url = {https://chrisvoncsefalvay.com/posts/lair},
  langid = {en-GB}
}
```

::: {.quarto-appendix-secondary-label}
For attribution, please cite this work as:
:::

::: {#ref-csefalvay2024 .csl-entry .quarto-appendix-citeas}
Chris von Csefalvay. 2024. "LAIR - Language As Intermediate
Representation." January 6, 2024.
<https://chrisvoncsefalvay.com/posts/lair>.
:::

</div>
:::
:::
