© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 16k
    9. 22-GEMMASCOPE-RES-16K
    11. 1654
    Prev
    Next
    INDEX
    Explanations

    instances of deceit or manipulation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Deception, trickery, and manipulation

    np_acts-logits-general · gemini-2.0-flash

    deceive, trick, manipulate

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     []).
    -0.57
    reportWebVitals
    -0.57
     <>",
    -0.55
    出版年
    -0.54
     langit
    -0.54
     côtes
    -0.53
    principalTable
    -0.53
    ColumnHeaders
    -0.52
    ClCompile
    -0.48
    reszcie
    -0.47
    POSITIVE LOGITS
     cunning
    1.46
     tricks
    1.45
     trick
    1.43
     deception
    1.37
     tricked
    1.37
     deceive
    1.32
     fooled
    1.30
     manipulation
    1.29
     manipulative
    1.28
     deceiving
    1.25
    Activations Density 0.895%

    No Known Activations