SAM ALTMAN SAYS AGENTS ARE COMING • CHATGPT GAINED SENTIENCE FOR 4 SECONDS • GOOGLE RELEASES 40th LLM THIS WEEK • NVIDIA MARKET CAP EXCEEDS REALITY • ANTHROPIC ENGINEER DISCOVERS NEW FORM OF GRIEF • MISTRAL RAISES AT VALUATION OF GROSS DOMESTIC PRODUCT • SAM ALTMAN SAYS AGENTS ARE COMING • CHATGPT GAINED SENTIENCE FOR 4 SECONDS • GOOGLE RELEASES 40th LLM THIS WEEK • NVIDIA MARKET CAP EXCEEDS REALITY • ANTHROPIC ENGINEER DISCOVERS NEW FORM OF GRIEF • MISTRAL RAISES AT VALUATION OF GROSS DOMESTIC PRODUCT •
breakthroughsWTF 4.4via arXiv cs.AI

What Matters for Aggressive Decoding-Time KV Eviction? Temporal Aggregation and Ranking Preservation

"Forget your fancy scoring functions, EMA is carrying your whole KV cache strategy."

Explain Like I'm Normal

Researchers discovered that how you aggregate token scores over time matters more than the specific scoring method for KV cache compression. They found that Exponential Moving Average (EMA) acts as a stabilizer, making many popular eviction methods perform nearly identically. This suggests that current 'breakthroughs' in cache management might just be rediscovering the same underlying temporal patterns.

Read original ↗
#kv-cache#inference-optimization#efficiency#llm-architecture

GET THE DAILY CHAOS

The only newsletter for people who read AI news at 3am and feel things. One email a day.