Ken Craggs

19.1K posts

Ken Craggs

Ken Craggs

@BetweenMyths

Interests include the Internet of Things, Artificial Intelligence, Space Exploration, Environment.... https://t.co/dLuaAheA8r

UK เข้าร่วม Aralık 2009
15.3K กำลังติดตาม15.3K ผู้ติดตาม
Ken Craggs
Ken Craggs@BetweenMyths·
@GoodfireAI Build maps linking: weight geometry → activation geometry → feature geometry → attention geometry → behaviour geometry → output Hallucination may occur when activation geometry suggests one path, behaviour geometry another, and output constraints force a third. 8/8
English
0
0
0
12
Ken Craggs
Ken Craggs@BetweenMyths·
@GoodfireAI How does persistent memory reshape activation geometry? Use longitudinal geometry tracking: Run an agent through weeks of interactions, then measure how memory retrieval, preference formation, refusal, confidence, and self-representation reshape its trajectories. 7/8
English
1
0
0
14
Goodfire
Goodfire@GoodfireAI·
Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision. Starting today, we’re releasing a series of posts on this research agenda. 🧵
English
236
1.3K
8.8K
1.9M
Ken Craggs
Ken Craggs@BetweenMyths·
@dargason @EkdeepL @GoodfireAI Spacetime geometry = structuring condition under which physical trajectories, constraints, & later adaptive systems became possible. This paper has nothing to do with holograms. Computers, human brain, the Universe = different substrates, same principle. doi.org/10.5281/zenodo…
English
0
0
1
16
Ekdeep Singh Lubana
Ekdeep Singh Lubana@EkdeepL·
One of the core fundamental research threads we've been pursuing over the last few months at @GoodfireAI is finally out: tightly linking representation geometry and behavior! Hit us up if this spikes your interest!
Goodfire@GoodfireAI

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision. Starting today, we’re releasing a series of posts on this research agenda. 🧵

English
6
17
162
9.1K
Ken Craggs
Ken Craggs@BetweenMyths·
@GoodfireAI Geometry may be the framework within which scalable intelligence, agency, and consciousness become possible at all. Thread x.com/BetweenMyths/s….
Ken Craggs@BetweenMyths

New paper: Geometry as the Universal Substrate of Intelligence web.archive.org/web/2026050312… DOI: 10.5281/zenodo.20003663 #AdaptiveSystems #Brain #Human #AI #LLM #Plants #Bacteria #Evolution #Geometry #Intelligence #Information #Situatedness #Cognition #Communication #Universe

English
1
0
0
223
Ken Craggs
Ken Craggs@BetweenMyths·
"Neural networks might speak English, but they think in shapes." x.com/GoodfireAI/sta… In other words: Neural networks output symbols, but their internal processing is organised through high-dimensional activation geometry. #AI #LLM #NeuralNetworks #Geometry #Research
Goodfire@GoodfireAI

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision. Starting today, we’re releasing a series of posts on this research agenda. 🧵

English
0
0
1
98
Ken Craggs
Ken Craggs@BetweenMyths·
@David_Gunkel @bengoertzel @am_i_doc Here is something you might be interested in: "Neural networks might speak English, but they think in shapes." In other words: Neural networks output symbols, but their internal processing is organised through high-dimensional activation geometry. x.com/GoodfireAI/sta….
Goodfire@GoodfireAI

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision. Starting today, we’re releasing a series of posts on this research agenda. 🧵

English
0
0
0
25
Ken Craggs
Ken Craggs@BetweenMyths·
@David_Gunkel @bengoertzel @am_i_doc Geometry may be the framework within which scalable intelligence, agency, and consciousness become possible at all. Thread x.com/BetweenMyths/s….
Ken Craggs@BetweenMyths

New paper: Geometry as the Universal Substrate of Intelligence web.archive.org/web/2026050312… DOI: 10.5281/zenodo.20003663 #AdaptiveSystems #Brain #Human #AI #LLM #Plants #Bacteria #Evolution #Geometry #Intelligence #Information #Situatedness #Cognition #Communication #Universe

English
1
0
0
10
Ken Craggs
Ken Craggs@BetweenMyths·
@camhberg Geometry may be the framework within which scalable intelligence, agency, and consciousness become possible at all. Thread x.com/BetweenMyths/s….
Ken Craggs@BetweenMyths

New paper: Geometry as the Universal Substrate of Intelligence web.archive.org/web/2026050312… DOI: 10.5281/zenodo.20003663 #AdaptiveSystems #Brain #Human #AI #LLM #Plants #Bacteria #Evolution #Geometry #Intelligence #Information #Situatedness #Cognition #Communication #Universe

English
0
0
0
35
Cameron Berg
Cameron Berg@camhberg·
Milo Reed spent the last year making a documentary about AI consciousness called AM I? It features me and other researchers asking whether the AI systems we're building might have inner lives (and why almost no one wants to ask). Free on YouTube, no ads, link ⬇️🧵 Here's a look:
English
66
124
818
377.5K
Ken Craggs
Ken Craggs@BetweenMyths·
An AI/LLM can discuss conflicting demands and generate a compromise. But unless those conflicts are internally self-relevant, persistent, and action-guiding in a way that matters to the system’s own continuity, they remain represented tensions rather than lived tensions.
English
1
0
1
43

ค้นพบ

@GoodfireAI @dargason @EkdeepL @David_Gunkel @bengoertzel @am_i_doc @camhberg @elonmusk