The Scaling Properties of Implicit Deductive Reasoning in Transformers

We investigate the scaling properties of implicit deductive reasoning over Horn clauses in depth-bounded Transformers. By systematically decorrelating provability from spurious features and enforcing algorithmic alignment, we find that in sufficiently deep models with a bidirectional prefix mask, implicit reasoning approaches explicit CoT performance across graph topologies and problem widths, though CoT remains necessary for depth extrapolation.

Paper

References (89)

Scroll for more · 38 remaining

Similar papers

© 2026 NYSGPT2525 LLC