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. 68855
    Prev
    Next
    INDEX
    Explanations

    the presence of whitespace and formatting elements in the text

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

    Japanese dialogue timestamps

    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
     sfeer
    -0.36
     veiligheid
    -0.32
    stdc
    -0.31
     Italijanski
    -0.31
    scul
    -0.30
     har
    -0.30
     Scul
    -0.29
    prudence
    -0.28
    styleUrls
    -0.28
     oorlog
    -0.28
    POSITIVE LOGITS
     asiatique
    0.59
     ویکی‌پدیای
    0.57
    endforeach
    0.53
     pegat
    0.50
     colonia
    0.49
    ⋱
    0.48
    principalColumn
    0.48
    EndGlobal
    0.48
    baomidou
    0.48
    Erstellt
    0.47
    Activations Density 0.002%

    No Known Activations