eTracer：通过声明级 grounding 实现可追溯的文本生成
eTracer: Polarity-Aware Evidence Grounding
AI总结：
eTracer 通过声明级 grounding 提升文本生成的可追溯性和可信度，改进了生物医学领域响应的验证效率。
摘要：
系统生成的响应如何高效验证，尤其是在高风险的生物医学领域？为解决这一挑战，我们引入了 eTracer，一个即插即用的框架，通过将声明与上下文证据进行 grounding 实现可追溯的文本生成。通过事后 grounding，每个响应声明都会与支持或反驳它的上下文证据对齐。基于声明级 grounding 结果，eTracer 不仅使用户能够精确地将响应追溯到其上下文来源，还量化了响应的忠实度，从而实现生成响应的可验证性和可信度。实验表明，我们的声明级 grounding 方法缓解了传统 grounding 方法在将生成语句与上下文句子级证据对齐时的局限性，从而在整体 grounding 质量和用户验证效率方面取得了显著提升。代码和数据可在 https://github.com/chubohao/eTracer 上获得。
英文摘要：
Evidence grounding makes generated content verifiable by linking statements to their sources. However, effective verifiability requires two properties that existing methods fail to jointly provide: verification granularity, the ability to localize fine-grained evidence for each statement, and verification completeness, the ability to capture both support and contradiction. Existing methods either operate at granularities too coarse for rapid inspection or are support-only. We formalize polarity-aware evidence grounding: linking each response sentence to its contextual evidence with a supporting, contradicting, or irrelevant label. We construct eTracer-Bench, a human-annotated benchmark with 7,905 supporting and contradicting sentence pairs across seven domains, covering both explicit and implicit contradictions. We further propose eTracer, a post-hoc framework whose context-aware signed scorer recovers the full evidence map in a single forward pass per response, with localization and polarity emerging jointly from one mechanism. eTracer achieves the best overall Citation F1 of 82.39, driven by a Contradict F1 of 89.20. A pre-registered within-subjects user study (n = 30) shows that polarity-aware grounding nearly doubles contradiction detection over the document-level baseline (27.9% to 51.9%, Cohen's d = 0.82, p = 0.0003), and suggests a false-completeness bias under which support-only grounding fails to improve over this baseline (23.4% vs. 27.9%). Dataset and code: https://github.com/chubohao/eTracer.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
