ACL 2026 Paper: LLMs Stick to Old Knowledge Despite New Information
mdredze · x · 2026-08-19
This paper presented at ACL 2026 finds that LLMs often continue to rely on information learned during pre-training, even when instructed to use new information. The study explores the update mechanisms and limitations when models face knowledge conflicts, revealing the difficulty in effectively overriding existing memory in RAG or in-context learning scenarios.
More from Models
- Ornith 1.5 35B A3B beats Flint on hallucination and overthinking in user tests — BarberIcy366 · 2026-10-23
- Meta researcher: horrifying infra, not talent, is why OSS can't catch closed frontier — A_K_Nain · 2026-10-23
- TypeSafe's Jev: a "Decision Model" LLM That Returns Floats at $0.042/M Input Tokens — ariG23498 · 2026-10-23
- Microsoft's MAI-Base-2 scaling ladder pegs the L42 run at ~1.8×10^23 FLOPs — teortaxesTex · 2026-10-23
- Local model agentic coding benchmark: 27B matches 125B, tool-calling is the lever — Joralk_ · 2026-10-23
- NVIDIA open-sources PersonaPlex-7B, a full-duplex voice AI model that listens and talks simultaneously — solyarisoftware · 2026-10-23