← Back to context Comment by llmslave2 1 month ago > But the capabilities of LLMs is clearly exponential and perhaps super exponentialBy what metric? 3 comments llmslave2 Reply aspenmartin 1 month ago - Scaling laws (Chinchilla type)- METR task horizonIt's a mix, performance gains are bursty but we have been getting a lot of bursts (RLVR, test-time compute, agentic breakthroughs) jennyholzer3 1 month ago Chat GPT told him it was true utopiah 1 month ago BS metric... /s
aspenmartin 1 month ago - Scaling laws (Chinchilla type)- METR task horizonIt's a mix, performance gains are bursty but we have been getting a lot of bursts (RLVR, test-time compute, agentic breakthroughs)
- Scaling laws (Chinchilla type)
- METR task horizon
It's a mix, performance gains are bursty but we have been getting a lot of bursts (RLVR, test-time compute, agentic breakthroughs)
Chat GPT told him it was true
BS metric... /s