specific method · filed under optimization
Anti-hallucination training
A training method described as reducing fabricated clauses and misquotes.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
The anti-hallucination training reduces fabricated clauses and misquotes.
usedtraining objectivein Hy3Tencent
Filed alongside
Other methods under optimization :: training stability.
Weight paddingActivation or logit clippingBehavioral regularizationCross-replica model-weight hash consistency checksElevated constant-learning-rate training stability stress testExponential moving average of checkpointsGradient clippingLearning-rate elevation for training stability stress testingLoss masking for excessively stale tokensOff-policy sample filteringPer-token regularization for off-policy RLSafeguards against training drift and reward hackingSoft droppingTraining stability stress testingWeight clippingWeight decay coupled to learning-rate squared