concept
active
concept:in-context-learning-icl

in-context learning (ICL)

Test-time adaptation from prompt or retrieved context with no parameter updates.

Neighborhood — ranked by edge-count

Frameworks (4)

framework

Claims (1)

claim

Methods (3)

method
  • Activation Steering
    associated_with
    Causal intervention technique: edit NLA explanation, reconstruct via AR, use difference as steering vector to manipulate model behavior.
  • Providing k labeled examples in the prompt to steer model behavior.
  • Method of using base models (no post-training) to observe spontaneous self-referential behaviors without confound of memorized introspection language.

Concepts (8)

concept
  • Induction Heads
    associated_withimplements
    Mechanistic circuits in transformers documented by Olsson et al. 2022, cited as evidence for pattern-repository assumption
  • semantic anchoring
    associated_with
    The central idea that external structure binds latent patterns to desired targets.
  • Transformers are recurrent through autoregression because the K/V stream provides horizontal information flow across positions, even though each forward pass is feedforward.
  • Skip-Trigram
    associated_with
    A three-token pattern of the form [source]...[destination][out] that one-layer attention heads implement; the paper's key characterization of one-layer transformer behavior
  • Belief Trajectory
    associated_with
    Sequence of belief states through a story.
  • Regime with dozens to hundreds of examples where sudden learning curves are seen.
  • Mesa-Optimizer
    associated_with
    A learned optimizer running inside a base optimizer; transformers proposed as mesa-optimizers implementing gradient descent in-context
  • Learning mechanism: parameter updates resemble classical Hebbian learning with associative and decay terms.

Artifacts (1)

artifact

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

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