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. 12651
    Prev
    Next
    INDEX
    Explanations

    expressions of emotional struggle and interpersonal conflict

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

    Following certain pronouns/articles

    np_acts-logits-general · gemini-2.0-flash

    predictable word sequences

    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
     élevées
    -0.76
    setVerticalGroup
    -0.67
     giustizia
    -0.62
    willReturn
    -0.60
    invokeLater
    -0.59
     barbati
    -0.59
    tagHelperRunner
    -0.58
     garanzia
    -0.58
     considér
    -0.58
     efficaces
    -0.57
    POSITIVE LOGITS
     [?]
    0.60
    letoe
    0.59
    orsese
    0.57
    neaux
    0.52
    koła
    0.52
    lankton
    0.51
     
    0.50
    inée
    0.50
     DRE
    0.49
    bourg
    0.49
    Activations Density 0.134%

    No Known Activations