A recurrent-depth architecture enables language models to improve reasoning performance by iterating computation in latent space, achieving gains equivalent to much larger models on benchmarks.
miniCTX : Neural theorem proving with (long-) contexts
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Non-reasoning LLMs fail the equivalence class problem while reasoning LLMs perform better but remain incomplete, with difficulty peaking at phase transition for the former and maximum diameter for the latter.
citing papers explorer
-
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
A recurrent-depth architecture enables language models to improve reasoning performance by iterating computation in latent space, achieving gains equivalent to much larger models on benchmarks.
-
How Well Do LLMs Perform on the Simplest Long-Chain Reasoning Tasks: An Empirical Study on the Equivalence Class Problem
Non-reasoning LLMs fail the equivalence class problem while reasoning LLMs perform better but remain incomplete, with difficulty peaking at phase transition for the former and maximum diameter for the latter.