Visualizing Graph-to-Answer Mechanism Recovery in Materials-Science Hypothesis Generation
Shashwat Sourav, Subhadeep Pal, Markus J. Buehler, Sanjay Das, Fiona Y. Wang, Dominik Soos, Tirthankar Ghosal
cs.CL, cs.AI
Submitted: 2026-08-04
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
The gist: AI co-scientists can generate fluent materials-science hypotheses, but fluency does not show that an answer preserves a scientifically meaningful mechanism.
Terminology
Abstract
AI co-scientists can generate fluent materials-science hypotheses, but fluency does not show that an answer preserves a scientifically meaningful mechanism. We present a graph-to-answer mechanism-tracing case study for Graph-PRefLexOR-8B, a Qwen3-8B model adapted to expose distinct stages for brainstorming, graph construction, pattern extraction, and synthesis. We organize semantic backtracking, graph corruption, activation-based recovery measurements, and layer-by-token-region grids into a visual diagnostic workflow for inspecting this pathway. Across 100 open-ended materials-science questions, final answers remain closest to the model's own structured stages, especially synthesis. Under graph corruption, a full sweep over 37 residual-stream checkpoints, the embedding output and 36 transformer blocks, shows little mechanism recovery in the earlier transition region at layers 7--10, recovery instead concentrates in late synthesis and answer-start regions around layers 30 and 36. The workflow is intended to help scientists and model developers identify where a generated hypothesis loses or regains mechanism support before it is passed to downstream experimental planning.
Sources
- Chain-of-Thought Unfaithfulness as Disguised Accuracy
- Hypothesis Generation for Materials Discovery and Design Using Goal-Driven and Constraint-Guided LLM Agents
- Measuring Faithfulness in Chain-of-Thought Reasoning
- Locating and Editing Factual Associations in GPT
- Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination
- Unifying Large Language Models and Knowledge Graphs: A Roadmap
- Is Cosine-Similarity of Embeddings Really About Similarity?
- Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting
- IterGen: Iterative Semantic-aware Structured LLM Generation with Backtracking
- Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small
- Tree of Thoughts: Deliberate Problem Solving with Large Language Models
- Towards Best Practices of Activation Patching in Language Models: Metrics and Methods
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