In 1-3B instruction-tuned LMs on GSM8K, arithmetic CoT readout is dominated by positional copying of the trailing number before the answer delimiter, accounting for 54-92 percentage points of accuracy.
Bogdan, Uzay Macar, Neel Nanda, and Arthur Conmy
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
representative citing papers
PluRule is a new multimodal multilingual benchmark showing that state-of-the-art vision-language models perform only marginally better than a trivial baseline at detecting specific rule violations in pluralistic online communities.
InsightReplay improves long CoT reasoning by extracting critical insights from the trace and replaying them near the active frontier, delivering +1.65 average accuracy gain across 24 model-benchmark settings.
LLMs settle on their answer after a minority of CoT tokens and produce an average 760 more as post-decision explanation, enabling early stopping that saves 500 tokens per query at a 2% accuracy cost.
SLRC quantifies genuine step necessity in LLM reasoning as a causal estimator, LC-CoSR training reduces rigidity with stability guarantees, and evaluations reveal a faithfulness-sycophancy paradox across frontier models.
citing papers explorer
-
The Readout Shortcut: Positional Number Copying Dominates Arithmetic CoT Readout in Small Language Models
In 1-3B instruction-tuned LMs on GSM8K, arithmetic CoT readout is dominated by positional copying of the trailing number before the answer delimiter, accounting for 54-92 percentage points of accuracy.
-
PluRule: A Benchmark for Moderating Pluralistic Communities on Social Media
PluRule is a new multimodal multilingual benchmark showing that state-of-the-art vision-language models perform only marginally better than a trivial baseline at detecting specific rule violations in pluralistic online communities.
-
Stateful Reasoning via Insight Replay
InsightReplay improves long CoT reasoning by extracting critical insights from the trace and replaying them near the active frontier, delivering +1.65 average accuracy gain across 24 model-benchmark settings.
-
Large Language Models Decide Early and Explain Later
LLMs settle on their answer after a minority of CoT tokens and produce an average 760 more as post-decision explanation, enabling early stopping that saves 500 tokens per query at a 2% accuracy cost.
-
Measuring and curing reasoning rigidity: from decorative chain-of-thought to genuine faithfulness
SLRC quantifies genuine step necessity in LLM reasoning as a causal estimator, LC-CoSR training reduces rigidity with stability guarantees, and evaluations reveal a faithfulness-sycophancy paradox across frontier models.
- Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thought