Tagged with

2 articles found

Abliteration: Performing Brain Surgery on LLMs to Cure Their Slop Addiction

How a controversial model surgery technique is cutting the flowery cruft from AI outputs, without retraining a single parameter.

#abliteration#llm-optimization#model-interpretability...

llm-optimization

DeepSeek OCR Flips the Script on Multimodal LLM Efficiency

DeepSeek’s new OCR model introduces a paradigm shift by making visual tokens more efficient than text tokens, challenging traditional assumptions in multimodal AI architecture.

#llm-optimization#moa#open-source-ai