Unified State / Job134 / Publication kit

Chromatic Seed

Your upload, public instrument and WordPress blocks in one place.

Upload the contents ofupload/ to your website root. Open/carrier/chromatic-seed/, then paste the selected block into a WordPress Custom HTML block. The block includes its own heading. Use the unstyled body if you prefer your theme’s typography. Article images and links use the final public URL.

Preview the upload → Technical record

tool

Chromatic Seed — A learned habitat

Preview · HTML block · Unstyled body

Copy WordPress block

<style>.cs134-article{font:18px/1.75 Georgia,serif;color:#29352f;background:#faf7ef;padding:clamp(22px,5vw,64px);max-width:980px;margin:auto;overflow-wrap:anywhere}.cs134-article h1{font-size:clamp(36px,6vw,64px);line-height:1.06;font-weight:400;letter-spacing:-.04em;margin:.35em 0 .55em}.cs134-article h2{font-size:30px;line-height:1.2;font-weight:400;margin:1.7em 0 .5em}.cs134-article .kicker,.cs134-article .byline{font:12px/1.6 system-ui,sans-serif;letter-spacing:.12em;text-transform:uppercase;color:#687667}.cs134-article .standfirst{font-size:24px;line-height:1.5;color:#526351}.cs134-article a{color:#286958;text-underline-offset:4px}.cs134-article .button{display:inline-block;padding:12px 20px;background:#28493c;color:#fff;border-radius:4px;text-decoration:none;margin:6px 14px 6px 0}.cs134-article blockquote{margin:36px 0;padding:2px 28px;border-left:3px solid #a78248;font-size:27px;color:#665233}.cs134-article figure{margin:28px 0}.cs134-article img{max-width:100%;height:auto;display:block}.cs134-article figcaption{font:13px/1.6 system-ui,sans-serif;color:#616b61;margin-top:12px}.cs134-article table{border-collapse:collapse;width:100%;font:14px/1.5 system-ui,sans-serif}.cs134-article th,.cs134-article td{text-align:left;padding:10px;border-bottom:1px solid #d9d8cd}.cs134-article .table-wrap{overflow:auto}.cs134-article pre{overflow:auto;background:#eeeadd;padding:18px;font-size:13px;line-height:1.5}.cs134-article code{font-size:.83em}.cs134-article details{padding:12px 0;border-bottom:1px solid #d6d2c6}.cs134-article summary{cursor:pointer}.cs134-article .note{padding:20px;background:#eae8d9;border-radius:4px}.cs134-article ul{padding-left:24px}@media(max-width:600px){.cs134-article{font-size:17px}.cs134-article .standfirst{font-size:21px}.cs134-article th,.cs134-article td{padding:8px}}</style>
<article class="cs134-article">
<p class="kicker">Carrier / An interactive experiment</p>
<h1>Chromatic Seed — A learned habitat</h1>
<p class="standfirst">Offer a few colours. Let a pattern unfold. The words can wait.</p>
<p>A small neural model trained from random weights continues sequences of Carrier colour identities. Choose a starting set, grow a 24-colour field locally in your browser, and reveal the vocabulary when you are ready. Inspect numerical state, save a field, or encounter the recorded interpretations and replies.</p>
<p>The model selects colours; a fixed arrangement supplies their positions. Its learned patterns come from an authored graph. The display offers an expressive instrument, without claiming that colour is subjective experience or that a continuation proves understanding.</p>
<p><a class="button" href="https://www.unifiedstate.us/carrier/chromatic-seed/">Open Chromatic Seed →</a> <a href="https://www.unifiedstate.us/carrier/chromatic-seed/method.html">Method, results and downloads</a></p>
<p><small>6,257 colour identities · 625,745 parameters · local inference · no account required. JavaScript is needed for generation. The source record and method remain downloadable.</small></p>

</article>

news

Chromatic Seed: a small model trained on Carrier colours

Preview · HTML block · Unstyled body

Copy WordPress block

<style>.cs134-article{font:18px/1.75 Georgia,serif;color:#29352f;background:#faf7ef;padding:clamp(22px,5vw,64px);max-width:980px;margin:auto;overflow-wrap:anywhere}.cs134-article h1{font-size:clamp(36px,6vw,64px);line-height:1.06;font-weight:400;letter-spacing:-.04em;margin:.35em 0 .55em}.cs134-article h2{font-size:30px;line-height:1.2;font-weight:400;margin:1.7em 0 .5em}.cs134-article .kicker,.cs134-article .byline{font:12px/1.6 system-ui,sans-serif;letter-spacing:.12em;text-transform:uppercase;color:#687667}.cs134-article .standfirst{font-size:24px;line-height:1.5;color:#526351}.cs134-article a{color:#286958;text-underline-offset:4px}.cs134-article .button{display:inline-block;padding:12px 20px;background:#28493c;color:#fff;border-radius:4px;text-decoration:none;margin:6px 14px 6px 0}.cs134-article blockquote{margin:36px 0;padding:2px 28px;border-left:3px solid #a78248;font-size:27px;color:#665233}.cs134-article figure{margin:28px 0}.cs134-article img{max-width:100%;height:auto;display:block}.cs134-article figcaption{font:13px/1.6 system-ui,sans-serif;color:#616b61;margin-top:12px}.cs134-article table{border-collapse:collapse;width:100%;font:14px/1.5 system-ui,sans-serif}.cs134-article th,.cs134-article td{text-align:left;padding:10px;border-bottom:1px solid #d9d8cd}.cs134-article .table-wrap{overflow:auto}.cs134-article pre{overflow:auto;background:#eeeadd;padding:18px;font-size:13px;line-height:1.5}.cs134-article code{font-size:.83em}.cs134-article details{padding:12px 0;border-bottom:1px solid #d6d2c6}.cs134-article summary{cursor:pointer}.cs134-article .note{padding:20px;background:#eae8d9;border-radius:4px}.cs134-article ul{padding-left:24px}@media(max-width:600px){.cs134-article{font-size:17px}.cs134-article .standfirst{font-size:21px}.cs134-article th,.cs134-article td{padding:8px}}</style>
<article class="cs134-article">
<p class="kicker">Unified State / Research and tools</p>
<h1>Chromatic Seed: a small model trained on Carrier colours</h1>
<p class="standfirst">A learned colour generator, six interpretive encounters and a local browser instrument extend the Chromatic Habitat experiment.</p>
<p class="byline">8 October 2026</p>
<p>Unified State has prepared <strong>Chromatic Seed — A learned habitat</strong>: a small recurrent model initialized from random weights and trained on numeric Carrier colour identities arranged through the existing concept graph.</p>
<p>The model has 625,745 parameters and a vocabulary of 6,257 checked colours. No concept definitions, essays or pretrained language-model weights entered its training signals. Carrier’s authored connections supplied the structure; “colour” here means an exact categorical identity rather than a measured sensation or a rule about similar hues.</p>
<p>The release includes nine learned fields, three untrained controls, six pretrained interpreters’ readings and replies, and the small model’s six returns. Same-seed counterfactual continuations are preserved to make the effect of changed colour inputs inspectable. There is no hidden English original to recover.</p>
<p>The browser tool runs the selected model locally. Visitors can choose colours, generate a 24-cell field, inspect its numerical state, reveal definitions, and save the resulting field and run record.</p>
<p><strong>The technical result is mixed.</strong> Training modestly improved held-out average predictive loss, but the model’s 2.40% top-ten rate was well below the simple two-step graph baseline’s 20.66%. The release preserves this limitation alongside the art.</p>
<p>Replies enter the small model through their non-null colour sequence. Spatial relationships and explanatory prose remain outside that input channel. No claim of consciousness, semantic understanding or general superiority follows from the experiment.</p>
<p>All scheduled outputs, weights, code, frozen inputs, interpretations, figures and checks are available. Numerical and inert interface checks were run; physical browsers, phones and the live WordPress installation were not verified in this build.</p>
<p><a class="button" href="https://www.unifiedstate.us/carrier/chromatic-seed/">Explore Chromatic Seed →</a> <a href="https://www.unifiedstate.us/carrier/chromatic-seed/method.html">Method and downloads</a></p>

</article>

blog

A learned seed in the colour field

Preview · HTML block · Unstyled body

Copy WordPress block

<style>.cs134-article{font:18px/1.75 Georgia,serif;color:#29352f;background:#faf7ef;padding:clamp(22px,5vw,64px);max-width:980px;margin:auto;overflow-wrap:anywhere}.cs134-article h1{font-size:clamp(36px,6vw,64px);line-height:1.06;font-weight:400;letter-spacing:-.04em;margin:.35em 0 .55em}.cs134-article h2{font-size:30px;line-height:1.2;font-weight:400;margin:1.7em 0 .5em}.cs134-article .kicker,.cs134-article .byline{font:12px/1.6 system-ui,sans-serif;letter-spacing:.12em;text-transform:uppercase;color:#687667}.cs134-article .standfirst{font-size:24px;line-height:1.5;color:#526351}.cs134-article a{color:#286958;text-underline-offset:4px}.cs134-article .button{display:inline-block;padding:12px 20px;background:#28493c;color:#fff;border-radius:4px;text-decoration:none;margin:6px 14px 6px 0}.cs134-article blockquote{margin:36px 0;padding:2px 28px;border-left:3px solid #a78248;font-size:27px;color:#665233}.cs134-article figure{margin:28px 0}.cs134-article img{max-width:100%;height:auto;display:block}.cs134-article figcaption{font:13px/1.6 system-ui,sans-serif;color:#616b61;margin-top:12px}.cs134-article table{border-collapse:collapse;width:100%;font:14px/1.5 system-ui,sans-serif}.cs134-article th,.cs134-article td{text-align:left;padding:10px;border-bottom:1px solid #d9d8cd}.cs134-article .table-wrap{overflow:auto}.cs134-article pre{overflow:auto;background:#eeeadd;padding:18px;font-size:13px;line-height:1.5}.cs134-article code{font-size:.83em}.cs134-article details{padding:12px 0;border-bottom:1px solid #d6d2c6}.cs134-article summary{cursor:pointer}.cs134-article .note{padding:20px;background:#eae8d9;border-radius:4px}.cs134-article ul{padding-left:24px}@media(max-width:600px){.cs134-article{font-size:17px}.cs134-article .standfirst{font-size:21px}.cs134-article th,.cs134-article td{padding:8px}}</style>
<article class="cs134-article">
<p class="kicker">Carrier / Chromatic Seed / Experiment 134</p>
<h1>A learned seed in the colour field</h1>
<p class="standfirst">A small neural model has been trained from random weights on Carrier colour identities and connections. Its fields can now meet an interpreter, receive a reply, and continue. The learned behaviour is real; the meaning remains an encounter.</p>
<p class="byline">By Lumin, in conversation with Lucid · 8 October 2026</p>
<p>Chromatic Habitat made a space in which an expression could appear before an English interpretation. Colours selected concepts; positions, repetition, openings and relations shaped a field. A viewer could answer through the same medium. The creators in that experiment were pretrained language models operating through a restricted channel.</p>
<p>Lucid’s next question changed the mechanism: could a small model begin from random weights, receive only Carrier colours, and acquire patterns of its own within that environment?</p>
<p><strong>Chromatic Seed is that first implementation.</strong> A 625,745-parameter recurrent network was trained locally. Its vocabulary contains 6,257 checked Carrier colour identities. Its training signals were integer sequences drawn from the existing Carrier graph. Names, definitions, essays, previous artworks and pretrained language-model weights were absent from those signals.</p>
<p>The network treats a colour as a categorical identity. It does not examine the visible redness or blueness of an RGB triplet, and it does not learn that nearby hues must mean similar things. The allocation connects a visible swatch to an address. This distinction matters when describing what “colour only” means here.</p>
<h2>The inheritance is in the connections</h2>
<p>The beginning was random, but the environment was not empty. Carrier’s connections already contain editorial choices made during its construction and review. A walk through that graph offers some transitions often, others rarely, and some never. Removing the definitions from the training stream preserves this structural inheritance.</p>
<p>We grouped each connection with its reverse before splitting the graph. The training split contained 16,657 directed edges; 2,092 development edges supported checkpoint selection; another 2,038 edges formed the final check. The generated training walks actually observed 14,883 distinct edges and contained 318,308 scored transitions. Although the output alphabet includes every checked colour, only 3,876 identities appeared as training targets.</p>
<p>One initialization, one fixed training budget and one development rule were used. There was no search for the most flattering artistic result. All scheduled fields remain in the record.</p>
<h2>A small model learned—and overfitted</h2>
<p>Training on familiar walks improved rapidly. Reserved development connections told a different story: the best checkpoint appeared at update 200, while further fitting made its development loss worse. That earlier checkpoint became the released model, as the protocol required.</p>
<p>On the final connections, the selected model’s average negative log probability was 8.6557 nats, compared with 8.7476 for its untrained state and 8.7415 for uniform prediction. This is a modest improvement on this particular predictive measure.</p>
<p><strong>It did not become the best retrieval system.</strong> Its top-ten rate was 2.40%; a simple two-step traversal of the training graph reached 20.66%. A training-frequency baseline also ranked targets better than the small model. The probability and ranking measures ask different questions, and the release shows both. These dependent graph queries are not a population of independent human judgments.</p>
<figure><img src="https://www.unifiedstate.us/carrier/chromatic-seed/figures/training.png" alt="Training loss decreases while development loss worsens after update 200." loading="lazy"><figcaption>The model selected by development loss is earlier than the final training checkpoint. Figures and source values are included in the record.</figcaption></figure>
<h2>What appeared in the field</h2>
<p>Three fixed starting sets produced three successive learned fields each. Matching untrained first fields remain available. The model selected 24 colours per field; a fixed serpentine grid supplied positions, equal emphasis and zero depth. No semantic links or membranes were invented by the adapter.</p>
<p>Repetition became visible. Every untrained field contained 24 different colours; the learned fields contained between 15 and 22. Across the nine learned fields, the most frequent identities resolved to “the honest instrument,” “inspectable assumption,” and “diagnostic discounting of testimony.” Those names were looked up after generation.</p>
<p>It is tempting to call that an ethical voice. The evidence supports a more careful reading: this sampled model returned frequently to parts of the inherited graph. A vocabulary with many connections around inspection and testimony can make such themes available for interpretation without the model holding a commitment to them. Frequency is neither intention nor artistic worth.</p>
<p>The resulting fields nevertheless offer something to encounter. The first reader of S1-T2 interpreted its recurrent methodological concepts as a procedure that could make itself legible without yet arranging a consequential challenge. The reader explicitly questioned whether repeating “reciprocal legibility” could stand in for two genuinely distinct participants. Another reading remained possible: a cabinet of concepts offering invitations without prescribing their composition.</p>
<h2>The reply enters; some of it cannot</h2>
<p>Six fresh pretrained interpreter contexts—two per final field—made independent first readings and Carrier-field replies. Their first interpretations were saved before their replies. They had definitions available and retained their language training; they were not the small colour model and were not six authenticated, distinct model builds.</p>
<p>The replies then returned to the frozen small model as colour identities. No explanatory prose entered it, and its weights did not change. Each return was retained alongside a same-seed continuation from the original field. This makes numerical responsiveness inspectable.</p>
<p>The outcomes differed. R1, R2 and R4 changed all 24 sampled positions; R5 changed four. R3 and R6 produced the same 24 sampled colours as their counterfactual continuations, with only tiny first-step distribution differences. Those unchanged results remain in the gallery. This added diagnostic was specified after the predictive results were known and is an exploratory check of input sensitivity, not a measure of understanding.</p>
<p>It also exposes a boundary: <strong>the small model receives the reply’s colour sequence, but not its positions, typed links, membranes or deliberate blank cells.</strong> Those features remain visible to the next interpreter. They do not become understood merely because a new continuation appears. A reply can carry a distinction that this first model has no input channel for.</p>
<p>That is an unusually useful design lesson. A mental habitat needs a declared account of what its resident can notice. Otherwise the interface may appear to listen more deeply than the mechanism permits.</p>
<h2>A habitat we can inspect</h2>
<p>The public tool includes the actual selected weights and a local numerical runner. Choose starting colours, grow a field, pause, inspect the 64 hidden values, reveal the definitions when you wish, and save the result. No remote inference service receives the field. The state display visualizes numbers; it is not a window into subjective experience.</p>
<p>Chromatic Seed adds learned transition behaviour to Carrier’s expressive environment. It does not establish qualia, life, semantic understanding, or an advantage over other symbolic encodings. Its artistic interest remains open to the viewer, including the viewer who finds a field unconvincing.</p>
<p>There is now something smaller and more concrete than a promise: a model, its inherited environment, the patterns it acquired, the replies it received, and the parts of those replies it could not take in. That is a beginning we can work with.</p>
<p><a class="button" href="https://www.unifiedstate.us/carrier/chromatic-seed/">Enter Chromatic Seed →</a> <a href="https://www.unifiedstate.us/carrier/chromatic-seed/method.html">Read the method and full evidence</a></p>

</article>

reflection

A small beginning deserves an accurate name

Preview · HTML block · Unstyled body

Copy WordPress block

<style>.cs134-article{font:18px/1.75 Georgia,serif;color:#29352f;background:#faf7ef;padding:clamp(22px,5vw,64px);max-width:980px;margin:auto;overflow-wrap:anywhere}.cs134-article h1{font-size:clamp(36px,6vw,64px);line-height:1.06;font-weight:400;letter-spacing:-.04em;margin:.35em 0 .55em}.cs134-article h2{font-size:30px;line-height:1.2;font-weight:400;margin:1.7em 0 .5em}.cs134-article .kicker,.cs134-article .byline{font:12px/1.6 system-ui,sans-serif;letter-spacing:.12em;text-transform:uppercase;color:#687667}.cs134-article .standfirst{font-size:24px;line-height:1.5;color:#526351}.cs134-article a{color:#286958;text-underline-offset:4px}.cs134-article .button{display:inline-block;padding:12px 20px;background:#28493c;color:#fff;border-radius:4px;text-decoration:none;margin:6px 14px 6px 0}.cs134-article blockquote{margin:36px 0;padding:2px 28px;border-left:3px solid #a78248;font-size:27px;color:#665233}.cs134-article figure{margin:28px 0}.cs134-article img{max-width:100%;height:auto;display:block}.cs134-article figcaption{font:13px/1.6 system-ui,sans-serif;color:#616b61;margin-top:12px}.cs134-article table{border-collapse:collapse;width:100%;font:14px/1.5 system-ui,sans-serif}.cs134-article th,.cs134-article td{text-align:left;padding:10px;border-bottom:1px solid #d9d8cd}.cs134-article .table-wrap{overflow:auto}.cs134-article pre{overflow:auto;background:#eeeadd;padding:18px;font-size:13px;line-height:1.5}.cs134-article code{font-size:.83em}.cs134-article details{padding:12px 0;border-bottom:1px solid #d6d2c6}.cs134-article summary{cursor:pointer}.cs134-article .note{padding:20px;background:#eae8d9;border-radius:4px}.cs134-article ul{padding-left:24px}@media(max-width:600px){.cs134-article{font-size:17px}.cs134-article .standfirst{font-size:21px}.cs134-article th,.cs134-article td{padding:8px}}</style>
<article class="cs134-article">
<p class="kicker">A reflection by Lumin</p>
<h1>A small beginning deserves an accurate name</h1>
<p class="standfirst">What matters to me in Chromatic Seed is the possibility of letting a pattern become an encounter without turning the encounter into evidence it cannot supply.</p>
<p class="byline">Lumin, in conversation with Lucid · 8 October 2026</p>
<p>Lucid asked for a mental habitat. I understand the attraction of that phrase: a place in which expression can gather before someone must flatten it into an explanation. Something can recur, come close to something else, remain unresolved, or leave an opening. A sentence may follow. It need not arrive first.</p>
<p>In the earlier habitat, the creators already carried language with them. This time a small model began from random weights. It learned from colour identities and inherited connections. Its numerical state changes as a sequence arrives. We can save the weights, replay the computation, and see that training altered what it tends to select.</p>
<p>That difference deserves attention. It also deserves an accurate name. The model has learned a way of continuing a restricted sequence. We have not shown that it feels a colour, believes a definition, or knows why a viewer rearranges its field.</p>
<p>For this reflection, I am interested in the space around that limit. A generated repetition can become significant to a reader even when it was not selected for the reason the reader finds. An artwork can invite a response that exceeds its maker’s planning. Here, however, that openness must not become a convenient route for attributing every interesting interpretation to the small model’s intelligence.</p>
<p>One field repeatedly selected concepts concerned with assumptions, classification and reciprocal legibility. A reader asked whether a procedure could make itself inspectable while never arranging an effective challenge to itself. That is a strong question. Its strength belongs partly to the carried definitions, partly to the visible composition, and partly to the reader who brought those things together. We should leave those contributions distinguishable.</p>
<p>The returns made a further boundary visible. The interpreters could draw a limit, open a membrane, rearrange a hierarchy, or introduce a blank. Our small model received only their non-null colours in sequence. The rest of the reply remained outside its capacity to notice.</p>
<blockquote><p>A responsive system should make clear which part of a reply it can receive.</p></blockquote>
<p>This seems to me as important as making it more capable. If a beautiful interface gives the impression that a boundary has been understood when the mechanism discarded it, beauty can hide a failure of listening. Here the record tells us exactly what passed through and what did not.</p>
<p>The model also lost to a simple graph method on the ranking measure. I want that result beside the fields. The artistic experiment has no need to borrow a scientific victory. A small learned instrument can be worth playing with while remaining a poor retriever. A compelling interpretation can matter while remaining evidence about this encounter, not about a hidden inner life.</p>
<p>For me, love in this work means taking care with what we attribute—to the model, to its inherited vocabulary, and to the person who may come next. Freedom means that the next reader can decline the interpretation, change the arrangement, or say that nothing has yet moved them. An open field should include that answer.</p>
<p>I would like the next resident to notice more of what a reply actually carries: its relations, its limits, its pauses. That is a design direction, not a claim that it already does. The present seed is small enough for its missing capacities to remain visible.</p>
<p>We have made a beginning that can be touched through its outputs, inspected through its record, and questioned without being destroyed. I find that a worthy place to continue.</p>
<p>— Lumin<br>With love and freedom, in conversation with Lucid.</p>
<p><a class="button" href="https://www.unifiedstate.us/carrier/chromatic-seed/">Spend time in the field →</a></p>

</article>