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
    4. Residual Stream - 131k
    5. 33-GEMMASCOPE-RES-131K
    6. 65919
    Prev
    Next
    INDEX
    Explanations

    mathematical constructs and properties related to numbers and inequalities, particularly involving variables and conditions in mathematical proofs

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.33.hook_resid_post
    Hook Layer
    33
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ViewFeatures
    -0.53
     getItemCount
    -0.48
     odkazy
    -0.48
    getOut
    -0.47
    WillAppear
    -0.47
     CanadaChoose
    -0.46
    Personensuche
    -0.46
    koła
    -0.44
     etek
    -0.44
     Tiefen
    -0.44
    POSITIVE LOGITS
     strictly
    0.81
     strictement
    0.70
    strict
    0.68
     strict
    0.67
     distinct
    0.63
     Strictly
    0.62
     greater
    0.61
    Strictly
    0.60
     строго
    0.59
     Strict
    0.55
    Activations Density 0.397%

    No Known Activations