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 - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 64723
    Prev
    Next
    INDEX
    Explanations

    phrases indicating direction or movement

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

    heading or going somewhere

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     parís
    -0.49
    BrowserModule
    -0.46
    genstein
    -0.46
    angliski
    -0.45
     Púb
    -0.44
     UITableViewCell
    -0.43
     comboBox
    -0.43
    igshid
    -0.42
     SwiftUI
    -0.42
    NameInMap
    -0.42
    POSITIVE LOGITS
     heading
    1.02
     headed
    0.91
     Heading
    0.90
    Heading
    0.86
    heading
    0.85
     rumbo
    0.70
     toward
    0.69
     towards
    0.69
     homeward
    0.69
     head
    0.67
    Activations Density 0.007%

    No Known Activations