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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 31-GEMMASCOPE-RES-131K
    6. 91177
    Prev
    Next
    INDEX
    Explanations

    phrases that describe being trapped or caught in various contexts

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

    trapped in

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Personendaten
    -0.90
    MigrationBuilder
    -0.61
     Infórmanos
    -0.60
    ConstraintMaker
    -0.60
    UnusedPrivate
    -0.59
    
    -0.59
     مرئيه
    -0.56
    SourceChecksum
    -0.54
     вікісторінку
    -0.53
    ioutil
    -0.53
    POSITIVE LOGITS
     bởi
    0.39
     inescapable
    0.36
     mediocrity
    0.36
    pozycji
    0.35
     misunder
    0.34
     monotony
    0.34
     misconceptions
    0.33
     bureaucracy
    0.33
     bruge
    0.32
     quicks
    0.32
    Activations Density 0.053%

    No Known Activations