
Global workspace theory
12
MENTIONS
2
EPISODES
2
PODCASTS
Search complete. 12 mentions across 2 episodes found for "Global workspace theory".
Sep 24, 2026
#215 - Nick Bostrom - "We Are Creating Superintelligence": The Existential Risk is Real
N
32:33Nick BostromGUEST
And there have been different theories proposed before for like accounting for the human case.
N
32:40Nick BostromGUEST
like if you have you give somebody some anesthetic or like you wonder like are they still conscious or not kind of matters if you're if you're just having amnesia and the surgery you don't remember the pain or whether you actually didn't experience the pain seems to be like something you wouldn't want to know and so there are different theories like the global workspace theory um attention schema theory higher order well so global workspace theory is the idea that there is a sort of global computational workspace there is like our brain has different modules and processes happening but there is a shared space where they can as it were send information that then gets broadcast to a lot of other modules and that being conscious of something means for that content that information content to be present in this global workspace roughly speaking So you could be doing unconscious processing of certain material.
N
33:31Nick BostromGUEST
We do a lot of that.
N
33:33Nick BostromGUEST
Maybe most of the information processing in the human brain is unconscious, but some of it gets kind of entered into this global stage in the brain, and then it's available to all these different other modules to affect that.
P
34:06Peter McCormackHOST
So consciousness itself was just an evolutionary construct for something that we require to advance as a living
N
34:14Nick BostromGUEST
thing.
N
34:15Nick BostromGUEST
Well, I mean, obviously evolution had something to do with producing these brains that then have the global workspace.
N
34:24Nick BostromGUEST
Interestingly, one can then look for analogs to this in AIs.
“WorkspaceBench: Evaluating Interpretability Methods for the Global Workspace” by camilablank, agam_bhatia, Euan Ong, Neel Nanda
T
0:00Type Three AudioNARRATOR
Workspace Bench Evaluating Interpretability Methods for the Global Workspace By Camilla Blank, Agam Bottia, Ewan Ong et al.
T
0:10Type Three AudioNARRATOR
Published on September 22, 2026.
T
0:16Type Three AudioNARRATOR
Heading TLDR There's a list of bullet points here.
T
0:21Type Three AudioNARRATOR
We introduce WorkspaceBench, a set of evaluations for how well an activation-to-text tool can read the contents of the global workspace of a model, that is the intermediate variables during a forward pass.
T
0:33Type Three AudioNARRATOR
The benchmark comprises 3,356 questions across 27 eval families, spanning topics in safety, logical reasoning, and multi-hop computation, with a subset for single-token output tools.
T
0:47Type Three AudioNARRATOR
A desirable property of good interpretability techniques is minimal hallucinations, so WorkspaceBench also provides a hallucination-focused eval.
T
1:36Type Three AudioNARRATOR
This is bad for COT monitorability and makes interpretability essential to actually understanding what is going on.
T
1:43Type Three AudioNARRATOR
A key goal of interpretability is to understand intermediate variables that a model uses to compute its answers.