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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 1321
    Prev
    Next
    INDEX
    Explanations

    phrases indicating physical movement or actions involving the verb "go."

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

    "To" followed by a verb like "visit" or "go."

    np_acts-logits-general · gemini-2.0-flash

    going to visit

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 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
     שוליים
    -0.76
     nakalista
    -0.69
    IntoConstraints
    -0.68
    VersionUID
    -0.67
    XmlAccessorType
    -0.65
    PerformLayout
    -0.63
    setVerticalGroup
    -0.63
    <unused28>
    -0.60
    <unused14>
    -0.60
    <unused41>
    -0.60
    POSITIVE LOGITS
     visit
    0.65
     visited
    0.53
     visiting
    0.49
     attend
    0.48
    visit
    0.48
     the
    0.48
     places
    0.47
     Visit
    0.46
    Visit
    0.46
     visits
    0.46
    Activations Density 0.267%

    No Known Activations