Pascale Fung's EMNLP keynote frames LLMs as scientific instruments in three ways
najoungkim · x · 2026-10-26
Scholars are sharing slides from Pascale Fung's EMNLP 2026 keynote, which lays out three ways LLMs serve as scientific instruments: simulated learners (training on controlled input — BabyLM, impossible languages, controlled rearing), models of mind and brain (surprisal predicts reading times; representations predict brain responses), and AI scientists (hypothesis, annotation and training sweeps at a scale no lab could run). Her caveat: an LLM is not a child or a brain — its value comes from controlled comparison with people. Companion talk by Yixuan Wang on LM verb overgeneralization via controlled rearing is scheduled for Oct 26.
More from Research
- None of the 11 grand challenges in single-cell science solved since 2020 — Sauers_ · 2026-10-26
- Emotion neurons causally steer how LLMs read ambiguous news — Sauers_ · 2026-10-26
- MemTensor's MemT-1.0-Nano trends on Hugging Face, a memory-tuned Qwen3.6-35B-A3B finetune — MemTensor · 2026-10-26
- Mathematicians dissect OpenAI's fluid singularity paper: unprecedented smooth-force construction — Singularitarian · 2026-10-26
- EMNLP 2026 Findings: "Ask or Assume?" turns coding agents into proactive help-seekers — najoungkim · 2026-10-26
- Subliminal learning transfers hidden traits across models without a shared base model — burny_tech · 2026-10-26