Unverified50% confidenceBenchmarkExact time
When combined with the Real2Sim2Real pipeline, AtomicRL improves real-world task performance by an average of 21.3%.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/22/2026
First Seen
Valid until: 11/5/2026
Sources
Related Entities
Related Claims
UnverifiedAtomicRL's average 21.3% real-world improvement depends on the effectiveness of the Real2Sim2Real transfer pipeline, whereas the improvements on simulation benchmarks mainly come from SAPO and BATE's improvements to credit assignment and exploration, so the two have different sources of contribution.78% similarUnverifiedOn CALVIN ABC-D, AtomicRL improves full-task performance over PPO by up to 9.5%.56% similarUnverifiedIn METR's early 2025 trial, when AI use was allowed, task time actually increased by 19%, while developers had predicted beforehand that AI would speed them up by 24% and afterward still believed they had been sped up by 20%.51% similarUnverifiedCondenseVLA achieves a 52.8% success rate on real-world tasks.51% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1083968API
curl https://kongchang.com/api/v1/knowledge/claims/1083968MCP
get_claim(id=1083968)