A text generation system has produced the following list of potential next words, ranked from most to least probable: ['happy', 'glad', 'joyful', 'cheerful', 'content', 'pleased']. If the system is configured to select the final output using a 'beam width' of 3, which set of words will be chosen?

Analyzing a Text Generation Output

A language model generates a ranked list of potential next words: ['run', 'walk', 'sprint', 'jog', 'move']. The final output selected from this list is ['run', 'walk']. The 'beam width' (K) used in this selection process must have been ____.