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. 85787
    Prev
    Next
    INDEX
    Explanations

    closing braces in code

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

    getter and setter methods

    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
    StoryboardSegue
    -1.03
     snippetHide
    -1.01
    ſchaft
    -0.96
    principalColumn
    -0.95
     الحره
    -0.94
     kasarigan
    -0.92
    SharedDtor
    -0.90
    EndInit
    -0.89
    PyExc
    -0.88
    twimg
    -0.86
    POSITIVE LOGITS
    ↵↵
    0.54
    ↵
    0.50
    <eos>
    0.41
    .
    0.35
     What
    0.35
    }
    0.35
     read
    0.35
     the
    0.35
     
    0.35
    bestimmungen
    0.34
    Activations Density 0.005%

    No Known Activations