TransformerLens called out: non-deterministic runs yet probes leave no trace
Sauers_ · x · 2026-10-19
In the thread on LLMs detecting activation reads, Sauers questions what TransformerLens is actually doing: it is literally non-deterministic from run to run, yet adding a probe doesn't change the activations. The discussion argues that a read-only operation should never affect the numerics, and that even with complex frameworks this side effect should be avoidable.
Related event: Inserting probes alters compiled Qwen3 logits, sparking debate(3 posts)→
More from Research
- solidcv: Rust/PTX/CUDA vision library claims 12.1x speedup over OpenCV — TheZachMueller · 2026-10-19
- Open leaderboard for non-neural AI launches with NNAI-Bench v0.1, 1,800 synthetic questions — Silver_Employ2617 · 2026-10-19
- Why Hardest Open Math Problems Stay Unsolved: Folklore Marks Them Inaccessible — IgorCarron · 2026-10-19
- scikit-learn Co-creator Gael Varoquaux to Present TabICL Tabular Foundation Model at Paper Club — GaelVaroquaux · 2026-10-19
- University of São Paulo surveys devs on LLM code generation and refactoring — EnzoRigazzoUsp · 2026-10-19
- mlpowered releases podcast deep-dive on what's inside LLMs — mlpowered · 2026-10-19