© 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. 42531
    Prev
    Next
    INDEX
    Explanations

    keywords and symbols commonly used in programming and data manipulation contexts

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

    literal strings and numbers

    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
    RectangleBorder
    -0.86
     ſche
    -0.65
    ſelf
    -0.65
     ſta
    -0.65
    AddTagHelper
    -0.63
    ſelves
    -0.61
    Xna
    -0.59
     ſte
    -0.58
     ſtre
    -0.57
    ніципалі
    -0.56
    POSITIVE LOGITS
    yes
    0.56
    string
    0.51
    true
    0.50
    hello
    0.50
     yes
    0.49
    Yes
    0.48
    test
    0.46
    Hello
    0.45
     Vereinigte
    0.44
     Yes
    0.44
    Activations Density 0.278%

    No Known Activations