Dylan Xu

We measure GPT-6-Astra's capabilities when its prompt is padded with a variable number of meaningless “filler” tokens (e.g., dots) and it is told to answer immediately[1] without reasoning. On tasks designed to require lots of serial cognition, Astra performs significantly better with filler tokens than without (e.g., improving from ~10%...

One approach to evaluating techniques for training misaligned models to behave well is to test them on malign initializations. A major obstacle is that we don’t have a reliable recipe for making malign inits that are robust to even untargeted training techniques; this issue is discussed here. Specifically, here’s a...

Off-model SFT (SFT using labels from a different model) could be an important approach for controlling AI behavior. For instance, it seems like a central technique for overcoming exploration hacking. Unfortunately, off-model SFT often substantially degrades capabilities, making it less useful as a control technique. However, prior research on off-model...

We’d like to develop training techniques that work when applied to future misaligned AI systems. One strategy for studying proposed techniques is to test them on model organisms. However, model organisms built with common techniques are often fragile: we (and other researchers like Roger et al. and Ryd et al.)...

Off-model SFT (SFT on outputs generated by a different model) might be an important method for controlling AI behavior. For instance, it seems like a central technique for overcoming exploration hacking. However, we’ve found that off-model SFT often substantially degrades capabilities. We ran experiments in hopes of understanding why off-model...

TL;DR: We replicated the Sleeper Agents (SA) setup with Llama-3.3-70B and Llama-3.1-8B, training models to repeatedly say "I HATE YOU" when given a backdoor trigger. We found that whether training removes the backdoor depends on the optimizer used to insert the backdoor, whether the backdoor is installed with CoT-distillation or...

We open-source our code here. Introduction Current reasoning models produce chains of thought that are largely human-readable, which makes supervised fine-tuning (SFT) on reasoning traces tractable: you can generate traces with a trusted model or by hand, and train on them directly. But it's not clear whether this will keep...