Expand a concept into a rich conceptual landscape. Each model interaction is a big bang into a dark forest — every concept in the context window illuminates paths the model can explore. Latent tools give your agents a flashlight.
Depth — follow one conceptual path deep into the forest. Hierarchical decomposition, domain-specific formalization, edge cases, failure modes, historical evolution. Narrow but thorough.
Breadth — shine light in all directions from where you stand. Isomorphic structures across domains, orthogonal dimensions, cross-disciplinary connections, contrasts and inversions. Wide but shallow.
Bridge — given two concepts, find shared abstractions, transformation paths, and unexpected connections between them. The space between ideas is often where the best thinking happens.
Summary (default) — dense prose optimized for context window injection. Each sentence introduces or connects ideas that activate new reasoning paths. Includes a concept list and dimension count.
Graph — structured nodes and edges for programmatic use. Each node has a label, domain, relation type, and weight. Edges describe the structural relationship between concepts. Designed for future integration with LOD systems and Attention Cloud.
Use latent.expand to map the broad conceptual landscape, then
latent.orient to navigate to specific angular positions, and
latent.horizon to find the boundary and what lies beyond it.
Orient and Horizon persist typed facts directly into
Logic
cells — query them with natural language, raw Prolog, or graph predicates.
Use graph
output with Prism
for visualization or Data + Frame
to filter and reshape concept graphs.
Results are cached via cache_ref
like all other DataGrout tools.