做可解释性研究的团队会发现，你依赖的 SAE 评估指标可能不可靠——TPP 和 SCR 已被证伪，建议改用 sae-probes 并关注新基准的进展。
一篇来自 arXiv 的论文对 SAEBench（稀疏自编码器标准评估套件）中的质量指标进行了审计，发现 Targeted Probe Perturbation (TPP) 和 Spurious Correlation Removal (SCR) 在标准设置下无法通过多种可靠性测试，不应再用于 SAE 评估。其他指标也存在噪声高、区分度低的问题。sae-probes 变体是测试中最可靠的指标，但仍难以区分同一架构的不同变体。研究结论指出，当前 SAE 领域需要更好的基准测试方法。
Are Sparse Autoencoder Benchmarks Reliable?
Sparse autoencoders (SAEs) are a core interpretability tool for large language models, and progress on SAE architectures depends on benchmarks that reliably distinguish better SAEs from worse ones. We audit the SAE quality metrics in SAEBench, the de-facto standard SAE evaluation suite, through three complementary lenses: reseed noise on a fixed SAE, ground-truth correlation on synthetic SAEs, and discriminability across training trajectories. We find that two of these metrics, Targeted Probe Perturbation (TPP) and Spurious Correlation Removal (SCR), fail multiple lenses at their canonical settings and should not be used to evaluate SAEs. The other metrics show higher reseed noise and lower discriminability than the field assumes. The sae-probes variant of $k$-sparse probing is the most reliable metric we tested, but even sae-probes struggles to separate variants of the same SAE architecture. Our results show the field needs better SAE benchmarks.