arXiv:2606. 14636v2 Announce Type: replace Abstract: Control-function instrumental-variable estimators pass an estimated first-stage residual to an outcome model.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.
arXiv:2606.14636v3 Announce Type: replace Abstract: Many two-stage estimators assess the first-stage learner by prediction error, even when the next stage uses its residual. In control-function instr...
arXiv:2606. 14636v1 Announce Type: new Abstract: Control-function instrumental variable estimators need a first-stage residual, not merely a first-stage prediction.
arXiv:2607. 25074v1 Announce Type: cross Abstract: Synthetic control (SC) matches a treated unit's pre-treatment trajectory to a weighted combination of donor units.
arXiv:2606. 18322v1 Announce Type: cross Abstract: Sparse Autoencoders (SAEs) decompose residual-stream activations into interpretable features.
ObserverBench is a benchmark framework that evaluates whether internal mechanistic estimators—called observers—are suitable for guiding interventions, control, or safety actions in language models. It separates estimation accuracy from the loss incurred by the chosen action, showing that accurate predictions do not always lead to better decisions. Experiments on GPT‑2‑small, Qwen2.5‑7B, Gemma‑2‑9B‑it, and Qwen3.5‑9B demonstrate that observers trained on action loss can reduce deployment loss, while traditional metrics like AUROC may rank monitors differently from actual performance.
arXiv:2609.23937v1 Announce Type: cross Abstract: Robust linear fits can resist response contamination yet remain too dense or unstable for useful global explanations. We propose penalized distillati...