Cameron Berg
AE Studio, AI Alignment and Consciousness Research
Functionalist What matters is the organisation of the processing, not the material
Cameron Berg is an AI alignment and interpretability researcher leading consciousness and self-representation investigations at AE Studio. His work examines how large-scale neural networks form internal representations of their own state, identity, and valence during pretraining and post-training alignment, using activation patching, sparse autoencoders, and linear probing to test whether functional precursors of self-awareness appear in transformer residual streams.
In September 2026, Berg co-authored The Pain Axis (arXiv:2609.16247) with Valen Tagliabue and philosopher Leonard Dung. Across 25 open-weight models spanning five model families, the study isolated a consistent linear direction encoding self-directed harm and showed that causal interventions along this axis systematically alter both first-person self-reports and downstream decision-making in trade-off environments.
On this site’s taxonomy, Berg holds the Functionalist position. His research program treats the organization of internal information routing, self-referential state tracking, and valence-guided action selection as empirically measurable properties in computational substrates, bridging mechanistic interpretability with philosophical indicator frameworks.
Known for. Mechanistic interpretability of self-referential and affective states in LLMs, co-author of The Pain Axis (arXiv:2609.16247)
Coverage on this site
Related researchers
- Leonard DungRuhr University Bochum, Center for Mind and CognitionSaving Artificial Minds (2026), indicator-based evaluation of animal and AI consciousness, co-author of The Pain Axis in large language models
- Valen TagliabueIndependent AI Interpretability and Welfare ResearcherFirst author of The Pain Axis (arXiv:2609.16247), linear representation probing and activation steering of self-directed harm in open-weight LLMs
- Hedayat AbedijooACE Conscious Studio / IndependentACE.await (2026), the ACE decision framework (Agency, Connection, Exchange)
-
Blaise Agüera y ArcasGoogle VP and Fellow, CTO of Technology & SocietyComputational functionalism, active inference, on-device neural computing, What Is Intelligence? (2025), and Who Are We Now? (2023)
- Igor AleksanderImperial College LondonAxiomatic Consciousness Theory, neural state machines, The World in My Mind, pioneering neural network engineering
-
Joscha BachCalifornia Institute for Machine ConsciousnessThe machine consciousness hypothesis, cyber animism, and MicroPsi
Cameron Berg is listed with the other functionalists and computationalists, who hold that what matters is the organisation of the processing, not the material. The theory family behind that position is set out in the index of consciousness theories. Every researcher covered on this site is indexed in the directory of consciousness researchers.