Amazon Mechanical Turk was used to annotate 1000 occurrences of each of 45 words for WordNet 3.1 word senses. All occurrences are drawn from MASC, except where MASC contains fewer than 1000 occurrences of a given word. In these cases, additional sentences were drawn from the Open American National Corpus (OANC).
The annotated words include a mix of nouns (n), verbs (v), and adjectives (j) drawn from texts in a range of genres. Each word was labeled by approximately 25 different annotators, for a total of roughly 1M total annotations.
The senses seen by the turkers for this annotation task are identical to the senses seen by trained MASC annotators for the MASC Sentence Corpus (described in Passonneau, Rebecca J.; Baker, Collin; Fellbaum, Christiane; Ide, Nancy. 2012. The MASC word sense sentence corpus. Proceedings of the 8th International Conference on Language Resources and Evaluation (LREC), pp. 3025-030. Istanbul, Turkey. May 23-25. ELRA.)
The MASC AMT Word Sense Annotations are available from the MASC Downloads page.