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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 24-GEMMASCOPE-RES-16K
    11. 4736
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate actions or strategies to evade certain situations or consequences

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

    "avoid" and its forms

    np_acts-logits-general · gemini-2.0-flash

    avoiding negative outcomes

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 24-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_24/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.24.hook_resid_post
    Hook Layer
    24
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ScopeManager
    -0.54
    ocument
    -0.52
     stringBuilder
    -0.52
    Grath
    -0.51
     redaktion
    -0.50
    TEntity
    -0.50
    ISupport
    -0.49
    oa̍t
    -0.49
    ModelSerializer
    -0.49
     enfans
    -0.48
    POSITIVE LOGITS
     AVOID
    0.82
     avoid
    0.81
    Avoid
    0.81
     avoidance
    0.78
    avoid
    0.78
     Avoid
    0.77
    Avoiding
    0.73
     avoided
    0.73
     avoiding
    0.68
     avoids
    0.66
    Activations Density 0.197%

    No Known Activations