Custom 'high' reasoning mode for 27B models shared on Reddit
A Reddit user shares a custom reasoning mode for 27B models, tested on vLLM, filling a gap between low and xhigh modes. The 'medium' mode reportedly erases explicit instructions, making the custom mode more effective.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Google AI's EnvHarness turns static agent benchmarks into adaptive training worlds
- AI agents that pass authentication can still drift, expose data, or get memory-poisoned
- Local video watermark remover released on CivitAI
- Video walks through implementing Kimi K3 from scratch in PyTorch
- SeedVR2 TensorRT Studio offers free open-source video upscaling