# Interpreting GFlowNets for Drug Discovery: What probes can and cannot show

arXiv: https://arxiv.org/abs/2511.19264 · HTML: https://mlsanity.com/paper/2511.19264/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 22 references that are in the map.
- First posted on arXiv 24 Nov 2025.

## Builds on (most cited first, 15 of 15)

- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Toy Models of Superposition](https://mlsanity.com/paper/2209.10652/) (2022)
- [GNNExplainer: Generating Explanations for Graph Neural Networks](https://mlsanity.com/paper/1903.03894/) (2019)
- [Network Dissection: Quantifying Interpretability of Deep Visual Representations](https://mlsanity.com/paper/1704.05796/) (2017)
- [Designing and Interpreting Probes with Control Tasks](https://mlsanity.com/paper/1909.03368/) (2019)
- [Scaling and evaluating sparse autoencoders](https://mlsanity.com/paper/2406.04093/) (2024)
- [Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation](https://mlsanity.com/paper/2106.04399/) (2021)
- [Algorithmic Recourse: from Counterfactual Explanations to Interventions](https://mlsanity.com/paper/2002.06278/) (2020)
- [Learning GFlowNets from partial episodes for improved convergence and stability](https://mlsanity.com/paper/2209.12782/) (2022)
- [BatchTopK Sparse Autoencoders](https://mlsanity.com/paper/2412.06410/) (2024)
- [CF-GNNExplainer: Counterfactual Explanations for Graph Neural Networks](https://mlsanity.com/paper/2102.03322/) (2021)
- [Drug discovery with explainable artificial intelligence](https://mlsanity.com/paper/2007.00523/) (2020)
- [A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models](https://mlsanity.com/paper/2503.05613/) (2025)
- [SynFlowNet: Design of Diverse and Novel Molecules with Synthesis Constraints](https://mlsanity.com/paper/2405.01155/) (2024)

## Abstract (arXiv)

Generative Flow Networks (GFlowNets) construct molecules through sequential decisions, but their internal policies remain opaque, limiting adoption in drug discovery, where chemists need interpretable rationales for proposed structures. We present a control-validated interpretability study of SynFlowNet, a synthesis-aware GFlowNet trained with a drug-likeness (QED) reward. Our framework combines gradient saliency and counterfactual edits, an undercomplete factor analysis, and an overcomplete BatchTopK sparse autoencoder, evaluated with shuffled-label controls, RDKit-descriptor baselines, a scaffold-disjoint split, cross-seed stability, and an architecture-matched untrained network. These controls materially change the interpretation. Physicochemical properties and functional groups are highly decodable from SynFlowNet embeddings, but an untrained network with the same architecture performs essentially as well as the trained policy (drug-likeness within 0.005 and marginally better molecular-size decoding). Thus, this decodability reflects graph architecture and atom featurization rather than representations acquired through policy training. The surviving conclusions are narrower: the overcomplete autoencoder reconstructs embeddings better than a matched undercomplete baseline at comparable sparsity; chemically enriched substructure detectors, including per-halogen and boron features, emerge in individual runs, although only a small subset of dictionary directions is stable across seeds; and zeroing individual features produces property-specific effects on probe decoding. Beyond SynFlowNet, this study provides a reusable control protocol for molecular-model interpretability, separating learned structure from signals supplied by architecture and input representation. High probe scores alone should not be treated as evidence of learned chemistry.