SP
BravenNow
Thinking in Different Spaces: Domain-Specific Latent Geometry Survives Cross-Architecture Translation
| USA | technology | ✓ Verified - arxiv.org

Thinking in Different Spaces: Domain-Specific Latent Geometry Survives Cross-Architecture Translation

📖 Full Retelling

arXiv:2603.20406v1 Announce Type: cross Abstract: We investigate whether independently trained language models converge to geometrically compatible latent representations, and whether this compatibility can be exploited to correct model behavior at inference time without any weight updates. We learn a linear projection matrix that maps activation vectors from a large teacher model into the coordinate system of a smaller student model, then intervene on the student's residual stream during gener

📚 Related People & Topics

Deep learning

Deep learning

Branch of machine learning

In machine learning, deep learning focuses on utilizing multilayered neural networks to perform tasks such as classification, regression, and representation learning. The field takes inspiration from biological neuroscience and revolves around stacking artificial neurons into layers and "training" t...

View Profile → Wikipedia ↗

Neural network

Structure in biology and artificial intelligence

A neural network is a group of interconnected units called neurons that send signals to one another. Neurons can be either biological cells or mathematical models. While individual neurons are simple, many of them together in a network can perform complex tasks.

View Profile → Wikipedia ↗

Entity Intersection Graph

Connections for Deep learning:

🌐 Explainable artificial intelligence 3 shared
🌐 Medical imaging 2 shared
🌐 Neural network 2 shared
🌐 Interpretability 2 shared
🌐 Artificial intelligence 2 shared
View full profile

Mentioned Entities

Deep learning

Deep learning

Branch of machine learning

Neural network

Structure in biology and artificial intelligence

}
Original Source
arXiv:2603.20406v1 Announce Type: cross Abstract: We investigate whether independently trained language models converge to geometrically compatible latent representations, and whether this compatibility can be exploited to correct model behavior at inference time without any weight updates. We learn a linear projection matrix that maps activation vectors from a large teacher model into the coordinate system of a smaller student model, then intervene on the student's residual stream during gener
Read full article at source

Source

arxiv.org

More from USA

News from Other Countries

🇬🇧 United Kingdom

🇺🇦 Ukraine