llama.cpp Users Report Infinite 'lalala' Loops with Gemma Models
Geritas · reddit · 2026-10-07
A Reddit user reports that for the past 3-4 days, llama.cpp has been producing broken output with Gemma4 31b: the model emits 'lalala' and loops indefinitely mid-conversation — a word that never appears in the scenario. With a screenshot attached, the user suspects a recent llama.cpp change and is looking for similar cases and possible causes.
More from Fun
- Viral quip: OpenAI convenes mathematicians, Anthropic convenes priests — SuB8u · 2026-10-07
- Dev reports OpenAI account banned right after failed CVP re-verification — dyn___ · 2026-10-07
- threepointone recommends krazam's 'You Are An Expert Software Engineer', likening it to qntm's 'The Difference' — threepointone · 2026-10-07
- Ethan Mollick mocks AI accounts pretending to grasp the latest math breakthroughs — emollick · 2026-10-07
- Mathematician admits OAI's Hodge conjecture progress outpaced expectations, mocks the hype — ctjlewis · 2026-10-07
- Stardock CEO mocks AI bubble talk: 'can we agree the computer thing is a bubble?' — draginol · 2026-10-07