Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 24496
    Prev
    Next
    INDEX
    Explanations

    syntactical or structural elements in code snippets

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

    Swift Protobuf

    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
     betweenstory
    -0.87
     purpoſe
    -0.78
    ſelf
    -0.75
     pleaſure
    -0.75
     Majefty
    -0.75
     Reſ
    -0.72
     Houſe
    -0.72
     ſtate
    -0.70
    EndInit
    -0.69
     Monfieur
    -0.69
    POSITIVE LOGITS
    :
    0.65
     a
    0.53
    :<
    0.48
     :
    0.48
    __:
    0.48
    ?:
    0.47
    ::<
    0.45
    ':
    0.45
    ():
    0.44
    :</
    0.43
    Activations Density 0.007%

    No Known Activations