Research Papers

Systematic explorations into the limits of test-time compute and algorithmic patience.

NOV 21, 2025

Scaling Test-Time Compute for Small Language Models: An Information-Theoretic Framework

ObscureAI Research Team

Recent advances in Large Language Models have demonstrated that performance scales predictably with model size and training data. However, for Small Language Models where parameter count is constrained, the primary avenue for performance gain lies in the effective utilization of inference-time computation.

Test-Time ComputeInformation TheoryScaling LawsSmall Language Models

Click to read full paper →

JAN 2, 2026

Faithfulness of Chain-of-Thought Explanations in Small Language Models

Anonymous Authors (Tsinghua University)

Chain-of-Thought (CoT) explanations are widely adopted as indicators of model reasoning in language models. Despite their popularity, the causal relationship between generated reasoning traces and final predictions remains unclear—particularly for small language models deployed on commodity hardware.

Chain-of-ThoughtFaithfulnessInterpretabilitySmall Language Models

Click to read full paper →