What actually happens inside the minds of language models?
How do LLMs represent knowledge, what circuits drive reasoning, and can we see their internal structure? Understanding the gap between external performance and internal mechanisms matters for safety and trust.
Language models develop human-like cognitive biases and reasoning patterns LLM internals reveal hidden trade-offs masked by performance metrics Synthesis layers organize research around thematic maps and extraction workflows. Language models show three hierarchical tiers of understanding with distinct mechanisms Steering a single latent feature triggers reasoning without chain-of-thought
How do knowledge layers differ functionally from reasoning layers in networks? How does LLM hallucination risk manifest in knowledge graph construction? How does the outer loop escape its own LLM's knowledge boundaries when discovering mechanisms? How do LLMs compress specific expert knowledge into median abstraction? Why does entity recognition act as a self-knowledge mechanism in LLMs? How do internal representations compare to human cognitive structures? Where do LLMs fail as knowledge systems compared to humans? What internal mechanisms explain LLM reasoning and representation limits? Which knowledge types do LLMs handle better than humans in reasoning tasks? What makes a neural network circuit actually interpretable to humans? Can LLMs have minimal introspection through causal linkage to internal states? How do knowledge and reasoning circuits interfere in the same neural network?See all 22 inquiring lines on this note →