Landau–Lifshitz scaling argument

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Deceptive capabilities may scale with model size (inverse scaling law hypothesis)hypothesis0.695
  Cited hypothesis from Lin et al. 2022 suggesting larger models become more capable of deception
• Scaling Laws for Activation Steering with Llama 2 Models and Refusal Mechanisms (Ali et al., 2025)concept0.693
  Related work finding larger models more resistant to steering, potentially consistent with ESR in 70B
• Introspective capacity may follow a simple monotonic scaling law across all concepts and architectureshypothesis0.690
  The paper treats this as possible but unconfirmed; current evidence shows concept-specific scaling only
• As LLMs scale, they develop increasingly general abstractions, with large models linearly representing abstract concepts like truth that capture shared properties of diverse inputsclaim0.688
  Interpretive claim connecting scale to abstraction level in LLM representations
• Landau theory of equilibrium phase transitionsframework0.688
  Theoretical framework for phase transitions; the paper builds on its scaling approach
• Free-energy scaling under domain-wall formation in Potts, autoregressive, and hierarchical networks shows that combinatorics of interactions on a graph prevent or allow spontaneous ordering.finding0.686
  Core result demonstrating topological constraints on self-organization
• The results generalise readily to non-equilibrium systems where scaling relationships remain similar (e.g., dynamic or localised scaling).claim0.683
  Claim about broader applicability of the scaling argument
• Scaling laws can be used to guide the training of sparse autoencoders.claim0.682
  Compute-optimal hyperparameters follow predictable power-law relationships.