Commonsense reasoning tasks S≈-2.15

Lower, more uniform anchoring for commonsense tasks

Source paper

extracted_from

(2025) · Edward Yi Chang · Kaya, Zeyneb N. · Ethan Chang

community

Few-shot anchoring & latent structure
members_of
How minimal examples disambiguate and recruit latent arithmetic/reasoning interpretations in LLMs
Anchoring bias in commonsense reasoning
members_of
Studies how LMs exhibit uniform anchoring effects (S ≈ −2.15) across commonsense tasks, decomposed by cohesion, mismatch, and budget forces.
Commonsense reasoning anchoring bias
members_of
Uniform anchoring effect quantified at S ≈ −2.15 across commonsense reasoning tasks

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.

Commonsense reasoning S ≈ -2.15 uniformfinding0.861
Lower, uniform anchoring for pattern-matching tasks.
Commonsense reasoning shows uniform but weaker anchoring (S ≈ −2.15)finding0.810
Task-specific comparison.
Commonsense tasks show weaker but uniform anchoring on LLaMA (S ≈ −2.15)finding0.809
E3 finding suggesting pattern matching requires less intensive processing than compositional reasoning
cyclic reasoning tasksconcept0.759
Language model reasoning tasks with cyclic geometric structure used to test manifold steering.
Math/code tasks S ≈ -1.65 at layers 8–12finding0.751
Task-specific peak anchoring score for structured reasoning domains.
Compositional Reasoningconcept0.747
Modular reasoning enabled by monoid and other algebraic laws; central to maintainability and correctness of functional systems.
Compositional reasoning beyond Aristotle?question0.733
Secondary research question examining whether compositional frameworks can transcend classical logic.
common senseconcept0.730
The practical, reality-based judgment that guides successful unfolding and adaptation.