Andrew Reece | PhD Behavioral Data Scientist

Principal Scientist at BetterUp Labs

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

Introducing the CANDOR corpus from @BetterUp: A 1TB, 850hr audio-video dataset of 1,656 unscripted conversations in America in 2020. CANDOR = Conversation: A Naturalistic Dataset of Online Recordings https://t.co/ww9LD34Qjm

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

@hwchase17 for Chat Agents using retrieval_with_sources tools, is it possible to store sources in memory? Intermediate steps don't save into history and Agent Final Answer keeps dropping source info. Thanks!

[Gus Cooney @guscooney]

CANDOR corpus is now available! It took years of hard work, but we hope it will be useful for researchers from many fields interested in conversation and social interaction. Dataset available for download (link in manuscript). Please share it widely.
https://t.co/2YELqLuC2I

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

In sum, if it's going to be robot overlords I at least hope it's the one who doesn't have to Google my pleas

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

I asked 5 hard questions to two modern language AIs - Meta's new BlenderBot (https://t.co/1jvt3JyfZT) and OpenAI's GPT-3. Warning: Gets a little computer nerdy, but there are riddles and metaphysics too!

Answers and commentary in :thread:

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

GPT-3 (first answer):
def square_root_of_sum(foo, bar, baz):
return math.sqrt(foo + bar + baz)

Blender: 0 (but why)
GPT-3: 5 (bad indentation forgiven)

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

@MeirSimchah Basically histograms of behaviors for good (blue) and bad (red) conversationalists. Y-axis = freq. a given behavior occurred, at different lvls of extremeness (x-axis bins). So high-right blue dot on vocal intensity means good Cs more frequently expressed high VI than bad Cs.👌

[Andrew Reece | PhD Behavioral Data Scientist @_andrewreece]

Introducing the CANDOR corpus from @BetterUp: A 1TB, 850hr audio-video dataset of 1,656 unscripted conversations in America in 2020. CANDOR = Conversation: A Naturalistic Dataset of Online Recordings https://t.co/ww9LD34Qjm