Tag
The paper proposes Stepwise-Querying, a method that elicits step-specific explanations from black-box teachers without accessing their internal chain-of-thought traces, assembling them into training data for supervised fine-tuning of smaller student models. It achieves the top score in 14 of 16 comparisons across two teachers, two students, and four math/code benchmarks, and introduces the SEIG diagnostic for measuring stepwise explanation information gain.
This paper introduces a framework for jailbreaking black-box LLMs through text-only continuation interfaces, combining sample-based distribution reconstruction, risk-gated residual control, and speculative multi-token execution to bypass safety alignment without access to model weights or token probabilities.
This paper proposes a model-agnostic, post-hoc interpretation framework for black-box LLMs using sentence-level energy landscapes. A surrogate Energy-Based Model simulates the target LLM, and a lightweight interpreter network identifies which prompt sentences most influence a given output, without requiring further API calls.