Erik Miehling IBM Research, Karthikeyan Natesan Ramamurthy International Business Machines, Praveen Venkateswaran International Business Machines, Ching-Yun Ko International Business Machines, Pierre Dognin International Business Machines, Moninder Singh International Business Machines, Tejaswini Pedapati, Avinash Balakrishnan International Business Machines 等 13 位作者
The AI Steerability 360 toolkit is an extensible, open-source Python library for steering LLMs. Steering abstractions are designed around four model control surfaces: input (modification of the prompt), structural (modification of the model’s weights or architecture), state (modification of the model’s activations and attentions), and output (modification of the decoding or generation process). Steering methods exert control on the model through a common interface, termed a steering pipeline, which additionally allows for the composition of multiple steering methods. Comprehensive evaluation and comparison of steering methods/pipelines is facilitated by use case classes (for defining tasks) and a benchmark class (for performance comparison on a given task). The functionality provided by the toolkit significantly lowers the barrier to developing and comprehensively evaluating steering methods. The toolkit is Hugging Face native and is released under an Apache 2.0 license at https://github.com/IBM/AISteer360.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations) · 2026 · 正式发表
Javier Carnerero-Cano Imperial College London, Universidad Carlos III de Madrid, Universidad Carlos II de Madrid and International Business Machines, Massimiliano Pronesti, Radu Marinescu International Business Machines, Tigran T. Tchrakian International Business Machines, James Barry, Jasmina Gajcin International Business Machines, Yufang Hou IT:U Interdisciplinary Transformation University Austria, Technische Universität Darmstadt and IBM Research Ireland, Alessandra Pascale International Business Machines 等 9 位作者
Large language models (LLMs) are widely used in knowledge-intensive applications but often generate factually incorrect responses. A promising approach to rectify these flaws is correcting LLMs using feedback. Therefore, in this paper, we introduce FactCorrector, a new post-hoc correction method that adapts across domains without retraining and leverages structured feedback about the factuality of the original response to generate a correction. To support rigorous evaluations of factuality correction methods, we also develop the VELI5 benchmark, a novel dataset containing systematically injected factual errors and ground-truth corrections. Experiments on VELI5 and several popular long-form factuality datasets show that the FactCorrector approach significantly improves factual precision while preserving relevance, outperforming strong baselines. We release our code at https://ibm.biz/factcorrector.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表