@rohanpaul_ai: LLMs can accept the same false claim differently depending on its tone, certainty, and grammatical form. Small wording …

X AI KOLs Timeline Papers

Summary

This paper presents EoBench, a benchmark testing LLMs' acceptance of false claims phrased in 19 different styles, finding that tone, certainty, and grammatical form significantly affect model responses, with larger and instruction-tuned models showing more resistance.

LLMs can accept the same false claim differently depending on its tone, certainty, and grammatical form. Small wording changes can make LLMs accept false claims, while larger and instruction-tuned models resist them more. Models must decide whether to trust a user’s new claim or rely on facts stored during training. EoBench tests this choice with about 66K false claims written in 19 styles across form, evidence, certainty, and tone. The team evaluated 18 Gemma, Llama, and Qwen models, then kept cases where each model already knew the correct fact. Commands, child-directed wording, formal language, and authority claims persuaded models most, while weak claims and counterfactuals persuaded them least. Across Llama and Gemma, larger models followed false context less often, and instruction tuning usually reduced that behavior. The finding shows that prompt wording can quietly change model answers, so evaluations and product safeguards must test linguistic framing directly. --- – arxiv. org/abs/2607.18232 Title: "It's Not What You Say, It's How You Say It: Evaluating LLM Responses to Expressions of Belief"
Original Article
View Cached Full Text

Cached at: 07/22/26, 06:23 AM

LLMs can accept the same false claim differently depending on its tone, certainty, and grammatical form.

Small wording changes can make LLMs accept false claims, while larger and instruction-tuned models resist them more.

Models must decide whether to trust a user’s new claim or rely on facts stored during training.

EoBench tests this choice with about 66K false claims written in 19 styles across form, evidence, certainty, and tone.

The team evaluated 18 Gemma, Llama, and Qwen models, then kept cases where each model already knew the correct fact.

Commands, child-directed wording, formal language, and authority claims persuaded models most, while weak claims and counterfactuals persuaded them least.

Across Llama and Gemma, larger models followed false context less often, and instruction tuning usually reduced that behavior.

The finding shows that prompt wording can quietly change model answers, so evaluations and product safeguards must test linguistic framing directly.


– arxiv. org/abs/2607.18232

Title: “It’s Not What You Say, It’s How You Say It: Evaluating LLM Responses to Expressions of Belief”

Similar Articles

Mind Your Tone: Does Tone Alter LLM Performance?

arXiv cs.AI

This paper investigates how tonal variations in prompts affect LLM accuracy on multiple-choice questions, finding systematic but model-dependent effects. The study uses multiple models and datasets to demonstrate that tone can significantly alter performance, cautioning against assuming tone-robust reliability.

Do LLMs Make More Mistakes If They Do Not Believe the Input Data?

arXiv cs.CL

This paper analyzes how LLMs' faithfulness to provided context depends on perceived plausibility, using factual, counterfactual, and fictional RDF triples in multiple languages. It finds a weak context–memory conflict and emphasizes that the choice of LLM judge can overestimate its strength.