© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 96649
    Prev
    Next
    INDEX
    Explanations

    braces and the structure of code blocks in programming languages

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

    code closing braces and statements

    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
    -------------</
    -0.54
    Dwight
    -0.45
    parsedMessage
    -0.44
    ⤹
    -0.44
    ceptives
    -0.43
    ligente
    -0.42
     Unidas
    -0.42
    iredo
    -0.41
    RegressionTest
    -0.41
    żesz
    -0.41
    POSITIVE LOGITS
    AutoScaleMode
    0.52
     ComVisible
    0.50
     defaultstate
    0.50
    ValueStyle
    0.48
    ">//
    0.45
     oprot
    0.41
    RTCK
    0.41
    InitVars
    0.39
     MaterialApp
    0.38
    appliquent
    0.35
    Activations Density 0.014%

    No Known Activations