Beyond Human-Like Processing: Large Language Models Perform Equivalently on Forward and Backward Scientific Text

Open in new window