Batch Effect
A phenomenon in which, during batched LLM inference, other inputs within the same batch indirectly influence the output distribution of the current input through the attention mechanism, causing the same input to yield different results under different batch compositions.
Timeline (last 90 days)
Domain shift refers to the distribution change caused by differences in experimental conditions, among which batch effect is the most common one.
Batch effects in biological data analysis refer to systematic biases introduced by different experimental batches or different sequencing platforms, which may make differences with no biological meaning appear significant.
All Facts (2)
Domain shift refers to the distribution change caused by differences in experimental conditions, among which batch effect is the most common one.
50%UnverifiedBatch effects in biological data analysis refer to systematic biases introduced by different experimental batches or different sequencing platforms, which may make differences with no biological meaning appear significant.
50%