Google Research: LLM factual errors are recall failures, not encoding gaps
dl_weekly · x · 2026-08-24
Google Research introduces a knowledge profiling framework distinguishing between fact encoding and recall. The study finds that frontier LLMs (like Gemini3 and GPT-5) encode nearly all facts but struggle to recall many. This suggests the factuality bottleneck is "lost keys" (recall failures) rather than "empty shelves" (encoding failures). Interventions should thus focus on post-training and inference-time retrieval methods.
More from Research
- LightSuite: fast cell detection and whole-brain atlas registration on a standard workstation — hisspikeness · 2026-10-15
- Pathway CSO pitches BDH as post-Transformer architecture uniting memory, reasoning, adaptation — schwentker · 2026-10-15
- AI-designed molecular scissors cut Alzheimer's proteins — but is it really AI? — gerardsans · 2026-10-15
- EmbodiedEvo tops RoboDojo at 38.95% success by letting robots inherit failed practice — thetripathi58 · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Google's HARNESS² evolves agent harnesses without verifiers, up to +11.5 points on WorkBuddy-Bench — dair_ai · 2026-10-15