most_similar not working as expected with GoogleNews Light Magnitude

[hrsmanian]

Sample snippet below.
Definitely most_similar to "help" is not "foster_feeling". Also, one would expect the similarity to be between -1.0 to 1.0. Seems like it is calculating eucledian distance

Am I doing something wrong or is this an issue?

from pymagnitude import Magnitude
wv = Magnitude("~/work/data/GoogleNews-vectors-negative300.magnitude.1", normalized = False)
wv.most_similar("help")
[('foster_feeling', 4.73423), ('eyewitness_accounts_background', 3.8858936), ('material_objectionable', 3.84861), ('Trail_rides_signify', 3.5484445), ('daring_TMV', 3.496218), ('free_Yahoo!_Account', 3.1454005), ('please_visit_http://www.TradeTheTrend.com', 2.9243984), ('Live_PR.com', 2.8446026), ('containing_inappropriate_links_obscenities', 2.7534983), ('Help_ICAL_CSV', 2.7428944)

Thanks
Ram