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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. MLP Out - 16k
    5. 0-GEMMASCOPE-MLP-16K
    6. 0
    Prev
    Next
    INDEX
    Explanations

    references to authority and leadership within the context of personality and beliefs

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

    terms related to authority and the influence of individual personalities

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

    formal, institutional language—especially abstract nouns and titles tied to legal or religious authority, often accompanied by concessive transitions like nonetheless or nevertheless

    oai_token-act-pair · gpt-5Triggered by @liuliujiujiu05
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-16k
    Configuration
    google/gemma-scope-2b-pt-mlp/layer_0/width_16k/average_l0_119
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.hook_mlp_out
    Hook Layer
    0
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    VersionUID
    -0.75
    DoubleQuotes
    -0.69
    WithTag
    -0.61
    بوابة
    -0.59
     CommandType
    -0.56
    InputTagHelper
    -0.52
    aktery
    -0.51
    μως
    -0.50
    ertas
    -0.50
    ρους
    -0.50
    POSITIVE LOGITS
    LookAnd
    0.70
     Commencez
    0.66
    WebElementEntity
    0.59
    Personensuche
    0.55
    Посилання
    0.55
    intios
    0.55
     AppCompatTheme
    0.54
    PYX
    0.53
     jLabel
    0.53
    feitura
    0.52
    Activations Density 1.410%

    No Known Activations