2 Commits

Author SHA1 Message Date
KS Jannette
362a47f88a Upgraded embedding model to voyage-3.5, updated README.md 2026-08-01 05:35:06 -04:00
KS Jannette
c6b07ccb56 fix minor css issue on stickies 2026-08-01 04:50:40 -04:00
5 changed files with 74 additions and 63 deletions

View File

@@ -8,11 +8,15 @@ In kongruity, the artifacts become "sticky notes." A board full of them looks ch
With a click, they are semantically evaluated, grouped into thematic clusters with descriptive headers, rankable and exportable to project planning and execution tools. With a click, they are semantically evaluated, grouped into thematic clusters with descriptive headers, rankable and exportable to project planning and execution tools.
## Voyage AI voyage-3.5
![Embedding model benchmarking.](Voyage.jpg)
## Clustering and evaluation: methodology ## Clustering and evaluation: methodology
Two models run in parallel, and neither sees the other's work. Anthropic's `claude-sonnet-5` (`backend/services/clustering.service.js`) reads the raw text of every note and groups them into labeled thematic clusters. Two models run in parallel, and neither sees the other's work. Anthropic's `claude-sonnet-5` (`backend/services/clustering.service.js`) reads the raw text of every note and groups them into labeled thematic clusters.
At the same time, Voyage AI's voyage-3 model (`backend/services/embedding.service.js`) converts each note's text into a numeric representation of its semantic meaning aka vector. At the same time, Voyage AI's voyage-3.5 model (`backend/services/embedding.service.js`) converts each note's text into a numeric representation of its semantic meaning aka vector.
Once the LLM returns, kongruity scores that grouping (`backend/services/validation.service.js`) using an established silhouette coefficient, with cosine distance rather than Euclidean as the proximity metric. Once the LLM returns, kongruity scores that grouping (`backend/services/validation.service.js`) using an established silhouette coefficient, with cosine distance rather than Euclidean as the proximity metric.

BIN
Voyage.jpg Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 114 KiB

View File

@@ -27,7 +27,7 @@ export const embedNotes = async (notes) => {
for await (const chunk of batches) { for await (const chunk of batches) {
const response = await client.embed({ const response = await client.embed({
input: chunk.map((n) => n.text), input: chunk.map((n) => n.text),
model: "voyage-3", model: "voyage-3.5",
}); });
response.data.forEach((item, i) => { response.data.forEach((item, i) => {

View File

@@ -95,7 +95,7 @@ const Stickies = () => {
<div className="clusters-container"> <div className="clusters-container">
{score != null && ( {score != null && (
<div className="cohesion-score"> <div className="cohesion-score">
Cluster cohesion: {scoreLabel(score)} Cluster cohesion: <strong>{scoreLabel(score)}</strong>
</div> </div>
)} )}
{rankedClusters.map((group, index) => ( {rankedClusters.map((group, index) => (

View File

@@ -40,10 +40,12 @@
} }
.cluster-header { .cluster-header {
position: relative;
display: flex; display: flex;
align-items: center; align-items: center;
gap: 12px; gap: 12px;
margin-bottom: 16px; margin-bottom: 16px;
min-height: 32px;
} }
.cluster-rank { .cluster-rank {
@@ -69,13 +71,18 @@
} }
.cluster-label { .cluster-label {
position: absolute;
left: 50%;
transform: translateX(-50%);
max-width: 50%;
margin: 0; margin: 0;
font-size: 1.2em; font-size: 1.2em;
font-weight: 600; font-weight: 600;
flex: 1; pointer-events: none;
} }
.cluster-drag-handle { .cluster-drag-handle {
margin-left: auto;
font-size: 1.4em; font-size: 1.4em;
color: #6dd6f4; color: #6dd6f4;
opacity: 0.4; opacity: 0.4;