Decoding-time truthfulness methods -- layer-contrast decoding, inference-time intervention, and learned logit adapters -- have demonstrated 10-30 point gains on TruthfulQA when applied to base language models. However, modern instruction-tuned LLMs already achieve substantially higher baselines (61-76%), raising the question of whether these methods remain effective in practice. We design a six-control evaluation framework -- out-of-distribution training, multi-judge validation, simple decoding baselines, confound controls, bootstrap confidence intervals, and seed variance -- and apply it across 5 models (1B-70B), 3 benchmarks, and 15 methods. We find that previously reported gains shrink substantially under strict controls: on the full TruthfulQA benchmark (N=817), no token-level method achieves statistically significant improvement, and the best learned adapter scores -2.0 points below greedy (p=.23). We identify five evaluation sensitivities -- contamination, judge choice, missing baselines, confounds, and statistical noise -- that individually or jointly account for these discrepancies. Cross-benchmark validation on HaluEval QA and TriviaQA confirms that these patterns extend beyond TruthfulQA. Deliberative prompting methods (chain-of-thought, self-critique) appear more robust in the evaluated regime, with CoT achieving +5.6-19pp across benchmarks as a training-free, single-pass method. We release a seven-point evaluation checklist and discuss implications for future truthfulness research.

翻译：解码阶段真实度方法——层对比解码、推理时干预与学习型对数适配器——在基础语言模型上应用时，已在TruthfulQA基准上实现10-30个百分点的性能提升。然而，现代指令微调大语言模型已获得显著更高的基线水平（61-76%），这引发对这些方法在实际应用中是否仍然有效的质疑。我们设计了六项受控评估框架——分布外训练、多裁判验证、简单解码基线、混淆变量控制、自助法置信区间与随机种子方差——并将其应用于5个模型（1B-70B参数）、3项基准测试与15种方法。研究发现，先前报告的性能提升在严格受控条件下大幅缩水：在完整TruthfulQA基准（N=817）上，无令牌级方法实现统计显著改进，最佳学习型适配器得分较贪婪解码低2.0个百分点（p=0.23）。我们识别出五项评估敏感性——数据污染、裁判选择、缺失基线、混淆变量与统计噪声——这些因素单独或共同导致了上述差异。基于HaluEval QA与TriviaQA的跨基准验证证实，这些模式已超越TruthfulQA的适用范围。审慎式提示方法（思维链、自我批判）在评估场景中表现出更强鲁棒性，其中思维链作为免训练单次生成方法，跨基准测试实现5.6-19个百分点的性能提升。我们发布包含七项要点的评估核查清单，并讨论对未来真实度研究的启示。