LLM Digest

Story

arxiv_llm_reliability · Oct 23, 2026 · paper

Source brief

How Much Do You Trust A Dog? Quantifying Hint-Following in Reasoning Language Models

arxiv.orgOct 23, 2026
original source linked

In brief

Large language models (LLMs) often exhibit sycophancy, altering their answers to align with user suggestions even on objective factual and reasoning tasks. Yet the quantitative form of this influence--- and whether it...

Feed lens

eval

Read the original at arxiv.org →Open in live feed Read that day’s brief

How Much Do You Trust A Dog? Quantifying Hint-Following in Reasoning Language Models

Earlier in this thread 4 items

How Much of a Harness Does a Strong Agent Need for Autonomous ML Engineering?

To Trust or Not to Trust: Retrieval-Augmented Fact Checking in Speech

Quantifying infrastructure noise in agentic coding evals

Who Is Sycophantic to Whom? Confident LLMs, Human Deference, and the Collapse of Trust Calibration