Brief

Researchers have developed a new method for stable randomized selection in competitive processes, such as funding or hiring. Their approach, termed the Clipped Linear Lottery, introduces a "smoothness" principle to ensure that minor score changes do not drastically alter selection probabilities. This method scales probabilities linearly between acceptance and rejection thresholds, offering a better tradeoff between stability and utility compared to existing lottery designs, as demonstrated by experiments on real-world peer review data. AI

Google Research has developed a framework to evaluate the alignment of Large Language Models (LLMs) with human behavioral dispositions, using established psychological assessments adapted into situational judgment tests. This approach quantizes model tendencies against human social inclinations, identifying deviations and areas for improvement in realistic scenarios. Separately, Google Research also introduced SLED (Self Logits Evolution Decoding), a novel method that enhances LLM factuality by utilizing all model layers during the decoding process, thereby reducing hallucinations without external data or fine-tuning. AI