© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 27407
    Prev
    Next
    INDEX
    Explanations

    trap or ambush tactics

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    adamkarvonen/qwen3-32b-saes/saes_Qwen_Qwen3-32B_batch_top_k/resid_post_layer_32
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    bul
    -0.10
    çī¦
    -0.09
    egis
    -0.09
    å·¯
    -0.09
    åıªèĥ½è¯´
    -0.09
     Pil
    -0.09
    å¼ĢåĬŀ
    -0.09
    ippo
    -0.08
    _PRI
    -0.08
    asaki
    -0.08
    POSITIVE LOGITS
    éĻ·éĺ±
    0.24
     trap
    0.21
     traps
    0.21
     Trap
    0.17
    trap
    0.16
    _trap
    0.16
    Trap
    0.15
    éĺ±
    0.14
     ambush
    0.13
     trapping
    0.13
    Activations Density 0.080%

    No Known Activations