Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 5512
    Prev
    Next
    INDEX
    Explanations

    proper nouns, particularly those related to geographical locations and character names in a narrative

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

    Tokens starting with "Lou", "Cou", "Hou", or "Bou"

    np_acts-logits-general · gemini-2.0-flash

    Lou, dou, cou, bou, hou

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    cipar
    -0.63
     consultato
    -0.56
    Lugar
    -0.54
    icali
    -0.51
    Walter
    -0.51
    tcs
    -0.50
    ects
    -0.49
    >>();
    -0.48
    ynos
    -0.48
    ']):
    -0.47
    POSITIVE LOGITS
     bou
    0.71
    hou
    0.70
     LOU
    0.70
     MOU
    0.68
    tvguidetime
    0.67
    cou
    0.67
    Hou
    0.67
    ագրություններ
    0.66
    strconv
    0.66
    LOU
    0.65
    Activations Density 0.122%

    No Known Activations