framework
active
framework:mixed-causal-model-m-l-one-hot-r-one-hot-p-gaussMixed causal model M (L one-hot, R one-hot, P Gauss)
Combined causal model estimating next-token distributions.
Neighborhood — ranked by edge-count
Papers (1)
paper
Thinkers (1)
thinker
- Theodora-Mara Pîslarassociated_withIntroduced combining causal models.
Concepts (1)
concept
- Causal abstractionextendsA framework the paper uses alongside feature geometry to deepen mechanistic understanding of LMs
Frameworks (3)
framework
- Prevailing view: retrieve entity by group position index.
- Retrieves target using the bound query entity.
- Retrieves entity via direct self-referential pointer.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Primary statistical model with random intercept by conversation, REML estimation, for pooled conversation-turn observations
- Formal representation of algorithms as directed acyclic graphs computing functions f_A
- Nonlinear models of brain dynamics that can be inverted via DEM.
- Generative models combining discrete and continuous states to link decisions and movements.
- The extent to which a probe direction, when intervened upon, actually changes model outputs — contrasted with mere classification accuracy
- Whether an internal direction causally controls a target behavior, verified by intervention success
- Antra's earlier model describing LM entity as three layers: base simulator, simulated simulator, simulated awareness; later revised as having less discreteness.
- Change point segmentation, kernel Dirichlet pooling, CV tuning.