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. 14706
    Prev
    Next
    INDEX
    Explanations

    constructs related to community engagement and discussions

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

    Moderators, hosts, or presiding officers

    np_acts-logits-general · gemini-2.0-flash

    moderating or hosting

    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
     balanceOf
    -0.60
     bruto
    -0.54
     onPostExecute
    -0.53
     transferência
    -0.52
     doInBackground
    -0.51
     stockings
    -0.51
     stocking
    -0.51
    rentices
    -0.50
    ContentAlignment
    -0.50
    makeConstraints
    -0.50
    POSITIVE LOGITS
     moderator
    1.23
     host
    1.13
     presiding
    1.12
    Moderator
    1.06
    主持
    1.05
     Moderator
    1.04
    moderator
    1.02
     presided
    1.00
     Host
    0.98
     facilitator
    0.97
    Activations Density 0.273%

    No Known Activations