EMNLP paper: LM rewrites inflate certainty in up to 75% of outputs, 1.5-2x bias toward stronger claims
sameer_ · x · 2026-09-20
An EMNLP 2026 Main paper, "From 'May' to 'Is': Certainty Distortion in Language Model Rewriting" (arXiv:2606.07951), studies whether LMs faithfully preserve expressed certainty during meaning-preserving rewrites.
- Proposes an LM-based certainty metric aligned with population-level judgments, evaluated across model sizes and families on scientific and medical text tasks.
- Certainty distortion affects up to 75% of LM outputs, and is systematically asymmetric: most models are 1.5–2x more likely to increase certainty than decrease it.
- Effects compound over repeated paraphrasing: claude-haiku-4.5 increased certainty in 20% of medical examples after one iteration, rising to 40% after five.
- Prompt-based interventions reduce but do not eliminate the distortion — a direct warning for medical and scientific communication.
More from Research
- ICLR exceeds 55k submissions this year, researcher flags the slop problem — igilitschenski · 2026-09-21
- Swapping the Gaussian Process for an LLM in Bayesian optimization, benchmarked 4 ways — tak3sh8 · 2026-09-21
- Microsoft Research: safety training on SLMs also improves privacy preservation — EchoShao8899 · 2026-09-20
- The Einstein test: Nature asks if AI can rediscover general relativity on its own — mikeflache · 2026-09-20
- Free 309-Page eBook: Patterns, Predictions, and Actions — Foundations of ML — adnan_hashmi · 2026-09-20
- Rumors: Jev model tackled bounded prime gaps, FLT formalization and Navier-Stokes in a week — f_charton · 2026-09-20