Temporal-probe decodability predicts reasoning in high-resource languages

measured in 1 paper

Bhatia et al. introduce MultiTempBench and evaluate 20 LLMs, combining a tokenization-fragmentation metric with linear-regression probes decoding Year/Month/Day (probe R^2 as temporal linearity) [bhatia-etal-2026-what-really-controls-temporal-reasoning-in-llms] A crossed mixed-effects regression over 285,000 predictions finds temporal linearity is the strongest predictor of reasoning accuracy in high-resource languages (English r=0.77) [bhatia-etal-2026-what-really-controls-temporal-reasoning-in-llms] Tokenization fragmentation dominates in low-resource languages (Hausa r=-0.97); the finding is purely correlational [bhatia-etal-2026-what-really-controls-temporal-reasoning-in-llms]

Structure

Linear Direction, Linear Separability

Context

temporal-reasoning, probe-geometry

Confirmed in models

Qwen3-0.6B · Qwen, Qwen3-4B · Qwen, Qwen3-14B · Qwen, DeepSeek-R1-Distill-Qwen-7B · DeepSeek, Llama-3.1-8B · Llama, Llama-2-7B · Llama, OLMo 3 7B · OLMo, Gemma 3 4B Instruct · Gemma, GPT-4o · GPT, GPT-OSS-20B · GPT-OSS

Method

Linear probing

Papers

What Really Controls Temporal Reasoning in LLMs: Tokenisation or Representation of Time? — Bhatia, Gagan, Isa, Ahmad Muhammad, Peyrard, Maxime, Zhao, Wei2026 · arXiv:2603.19017