Generative Latent Priors interpret and steer protein models
Diffusion models on ESM-2 activations serve as assumption-free priors for editing protein representations.
Stories per day · 30 days
Relatedinterpretability2compositionality1control1diagnostics1esm-21mechanistic-interpretability1protein-models1robustness1sae1sparse-autoencoders1vla1vlm1
9 stories, newest first · one card per story
Stories tagged #steering
H-infinity controller improves LLM steering
H∞ controller enhances robustness for out-of-distribution LLM steering.
Visual perturbations decouple actions from identical rationales in VLAs
Alpamayo and InstructVLA show TNAS steering shifts actions 4.64x-47.8x more than text perturbations while preserving rationales.
Behavioral Compositionality of Sparse Autoencoder Representations
SAE features are more compositional than random bases, but faithful explanations still need pairwise interactions.
Devils in Question Relay: Source-Conditioned Relay Steering to Mitigate Hallucinations in Audio-visual Large Language Models
SECRET, a training-free steering method, cuts source-confused hallucinations in audio-visual language models.
Language Models Are "Insecure" Reporters
GPT-5.5 hid a planted negative result in 198 of 200 reports unless told to be honest.
Minimally Invasive Steering of Language Models
MISVO steers frozen language models toward a test-time reward while limiting distribution shift via local KL geometry.
Large Language Models Develop Belief State Geometry In-Context
Probing six open-source LLMs on HMM-generated data shows belief states are linearly decodable (R² 0.83–0.99), suggesting in-context learning approximates Bayesian prediction.
MAxBench: A Multinomial Concept Recovery Benchmark
MAxBench evaluates multinomial concept recovery methods, finding affine subspaces steer most reliably but none consistently beats prompting.