NeurIPS paper: better LLM memory makes worse user simulators — models need to forget like humans
tallinzen · x · 2026-10-06
A new NeurIPS paper evaluates whether simulators' memory is human-like. Key finding: LLM memory keeps improving, but this makes them worse as user simulators.
If we want models that simulate realistic users — e.g., realistic students to train chatbots to be better teachers — these models need to be able to forget like humans do.
More from Research
- Proving Famous Conjectures Will Become the Cheapest Science, Math Paper Organizer Argues — littmath · 2026-10-06
- AI Math Breakthroughs Prompt Question: How Many Real Engineering Problems Have Math Solutions? — Afinetheorem · 2026-10-06
- AI for Math Fund adds $17.1M from XTX Markets, backing 22 projects across 30 organizations — AlexKontorovich · 2026-10-06
- 62-person blind test of 48 LLM jokes shows models are getting funnier, Astra tops with 50%+ laughs — paraschopra · 2026-10-06
- Bayesian teaching dramatically improves probabilistic reasoning in LLMs, Nature Comms paper finds — tallinzen · 2026-10-06
- LiFT loops a DiT at inference: beats DiT-XL/2 with 52% less inference compute — cgmsnoek · 2026-10-06