Bridging Linguistic Structure and Mechanistic Interpretability for Conceptual Interpretation in Language Models

arXiv:2408.11827 · cs.CL · Submitted 2026-08-24 · Read on arXiv

Nura Aljaafari, Danilo S. Carvalho, André Freitas

cs.CL

Submitted: 2026-08-24

Updated: 2026-08-25

Code: https://github.com/kingoflolz/mesh-transformer-jax

Terminology

• Transformer — 31× in this paper · explained in 2 episodes
• Trace — 22× in this paper · explained in 2 episodes
• Interpretability — 15× in this paper · explained in 15 episodes
• Abstraction — 13× in this paper · explained in 1 episode
• Aggregation — 11× in this paper · explained in 1 episode
• Activation Patching — 6× in this paper · explained in 2 episodes
• Definitional Semantic Roles — 5× in this paper
• Accuracy — 4× in this paper · explained in 1 episode
• Embeddings — 4× in this paper · explained in 2 episodes
• Mechanistic interpretability — 4× in this paper · explained in 2 episodes
• Transformer Models — 4× in this paper · explained in 2 episodes
• Adaptation — 3× in this paper · explained in 1 episode
• Multi-Layer Perceptron — 3× in this paper
• Alignment — 2× in this paper · explained in 1 episode
• Alternation — 2× in this paper · explained in 1 episode
• Consistency — 2× in this paper · explained in 3 episodes
• Feed-Forward Network — 2× in this paper
• Large Language Models (LLMs) — 2× in this paper · explained in 6 episodes
• Token Integration Mechanism — 2× in this paper
• alpha — 2× in this paper · explained in 1 episode
• AI Alignment — 1× in this paper · explained in 2 episodes
• Adaptability — 1× in this paper · explained in 1 episode
• Argument Structure Theory — 1× in this paper
• Clustering — 1× in this paper · explained in 2 episodes
• Explainability — 1× in this paper · explained in 3 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• Grounding — 1× in this paper · explained in 3 episodes
• In-Context Learning — 1× in this paper · explained in 4 episodes
• Knowledge Base Construction — 1× in this paper · explained in 2 episodes
• Latent Space — 1× in this paper · explained in 7 episodes
• Low-Rank Adaptation (LoRA) — 1× in this paper · explained in 9 episodes
• Multi-Head Attention — 1× in this paper
• Perplexity — 1× in this paper · explained in 3 episodes
• Pragmatics — 1× in this paper · explained in 2 episodes
• Prompt Engineering — 1× in this paper · explained in 4 episodes
• Rank-One Model Editing — 1× in this paper
• Semantic Similarity — 1× in this paper · explained in 2 episodes
• Transformer Architecture — 1× in this paper · explained in 2 episodes

Sources

• Discovering Knowledge-Critical Subnetworks in Pretrained Language Models
• Eliciting Latent Predictions from Transformers with the Tuned Lens
• Interpreting Neural Networks through the Polytope Lens
• Toy Models of Superposition
• The Pile: An 800GB Dataset of Diverse Text for Language Modeling
• How to use and interpret activation patching
• In-context Learning and Induction Heads
• LoRA: Low-Rank Adaptation of Large Language Models
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned
• Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering