The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology
2026-07-01 • Machine Learning
Machine Learning
AI summaryⓘ
The authors studied different ways to create special language models (called Model Organisms or MOs) used to test how well interpretability methods can find hidden behaviors. They found that how these MOs are trained greatly affects how easy it is to interpret their behaviors. Specifically, more realistic training methods often make MOs harder to interpret than the usual post-hoc approaches. Their work suggests that current MOs might not be very reliable for testing interpretability methods as previously thought.
Model Organisms (MOs)interpretabilitypost-hoc supervised fine-tuning (SFT)distributional policy optimization (DPO)activation oracleslogit lenssparse autoencoderslanguage modelstraining objectives
Authors
Andrzej Szablewski, Gabriel Konar-Steenberg, Raffaello Fornasiere, Nikita Menon, Stefan Heimersheim
Abstract
Model organisms (MOs) - language models trained to exhibit undesired or unnatural behaviours - are frequently used as testbeds for evaluating white-box interpretability techniques. Current MOs are typically constructed via post-hoc supervised fine-tuning (SFT) on behavioural transcripts or synthetic documents. Prior research has shown that interpretability methods can easily identify hidden behaviours in these MOs. However, recent work suggests that such post-hoc training methods may make interpretability unrealistically easy. We investigate this claim by constructing a suite of 54 $\verb|OLMo2-1B|$- and $\verb|gemma-3-1b-it|$-based MOs trained with seven different techniques, including standard post-hoc SFT, post-hoc DPO, and more realistic integration of MO data into the OLMo post-training DPO phase. We use these MO variants to benchmark activation oracles, activation steering, logit lens, and sparse autoencoders. Our findings show that (i) MO interpretability depends strongly on training objective, target behaviour, model architecture, and training data generation pipeline; (ii) substantial variance remains even after controlling for differences in the strength of target behaviour expression; and (iii) our more realistic $\textit{integrated training}$ often yields less interpretable MOs than standard post-hoc methods. Our results cast substantial doubt on the validity of current MOs as interpretability proxies.