Unverified50% confidenceBenchmarkExact time
In the Meta-World ML10/45 test environments, METTA achieves over 50% success rate on unseen tasks while maintaining a low training budget and short runtime.
1
Sources
50%
Confidence
Long-term
Relevance
10/23/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder naive training, the network is first trained to 100% accuracy on Task A, then trained only on Task B data, after which Task A accuracy drops from 100% to 64% while Task B reaches 100%.65% similarUnverifiedIn CLM training, pretraining alone scores 52% on the hard negative sample test; adding the hard negative sample stage jumps it to 69%; training with hard negative samples from day one peaks at only 62% and overfits quickly; omitting replay causes the hard negative sample score to drop from 69% back to 56%.64% similarUnverifiedIn reduced experiments using only 20%, 40%, and 80% of the training data, the performance gains persist stably, confirming the reliability of this method in low-resource settings.64% similarUnverifiedThe loss during training is calculated per token, and long samples contribute far more gradient than short samples, so a 10% share of rows can absolutely correspond to more than 50% of the token volume and gradient signal.63% similarUnverifiedIn the experiments, the 4B model's accuracy before training was about fifty percent, while the 27B model reached over seventy-five percent.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1088287API
curl https://kongchang.com/api/v1/knowledge/claims/1088287MCP
get_claim(id=1088287)