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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 18-GEMMASCOPE-RES-16K
    6. 3158
    Prev
    Next
    INDEX
    Explanations

    phrases related to making a positive first impression

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

    first impressions or impact

    np_acts-logits-general · gemini-2.0-flash

    making a good impression

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    stdc
    -0.53
    日閲覧
    -0.53
     Sargent
    -0.46
     ngh
    -0.44
     lustig
    -0.42
    ctivité
    -0.42
     experimentally
    -0.42
     Ol
    -0.42
    ::_('
    -0.42
    ENTY
    -0.42
    POSITIVE LOGITS
     impression
    1.87
     impressions
    1.70
     impress
    1.67
     impressed
    1.65
    impression
    1.60
     Impression
    1.59
    Impression
    1.51
     Impressions
    1.39
     impresion
    1.37
     indruk
    1.36
    Activations Density 0.179%

    No Known Activations