Activation Control Paper Method Review
How the Activation Controllability Benchmark paper works: a fixed copy task, a 0.85 similarity keep-rule, and residual-stream concept projections.
Gao’s logit-diff ablation is not a circuit rule
Gao et al.'s logit-diff ablation score measures how sparse an SAE latent's effect on output logits is. It is not a rule for admitting circuit components.
Gao’s citation gloss of the BERT illusion
Gao et al. (2024) cite the BERT interpretability illusion as a gloss: random activating examples can make an overly broad feature explanation look right.
How does arXiv:2412.09565v2 define obfuscated activations against a named latent-space monitor?
In arXiv:2412.09565v2 (Bailey et al., 8 Feb 2025), an activation is obfuscated against a named latent-space monitor D, not obfuscated in the abstract.
Vi-CD and WeightLens/CircuitLens: Which Paper-Stated Controls Make an Extracted Circuit Causal, and What Stays Unknown
An extracted subgraph is not yet a causal circuit. Read it as causal only after the paper’s own intervention controls are fixed, and do not fill parameters…
ETA vs. Li et al.: Interference Prefix, Response-Prefix and Prompt-Suffix Placement Are Three Different Procedures
These are three different early-token procedures, not three placements of one prefix. In Ding, Li, and Zhang, ETA ( arXiv:2410.06625v2 , ICLR 2025; version…
ACC++ (arXiv:2602.13483v2): The Exact Steps for Rank-1 QK Singular-Direction Candidates and Post-Softmax Removal
In arXiv:2602.13483v2, ACC++ forms one rank-1 candidate for each pair (upstream component, singular direction of the head’s QK matrix Ω), builds the…
Huang et al. (arXiv:2608.11295v1): Do J-Lens Traces Show Residual Trigger Awareness After Qwen3-8B Decontamination?
In Huang et al., arXiv:2608.11295v1, the post- </think> probe token, the 64–160-example backdoor and decontamination corpus, and the top-10 layerwise…
GPT-4 Wrote Explanations for 307,200 GPT-2 XL Neurons and Graded Them by Simulation
GPT-4 writes natural-language explanations of 307,200 GPT-2 XL MLP neurons across 48 layers, then scores them by simulating activations on held-out text.
Inner Workspace Talk Describes Residual Bottlenecks, Not a Global Mind
Read workspace talk as residual bottlenecks, not a psychological workspace. Each layer writes a d_model update; softmax does many-to-few selection.