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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13060
    Prev
    Next
    INDEX
    Explanations

    captions or descriptions of figures, charts, or images in scientific or technical documents.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @myinzi123

    references to specific categories or classifications in academic or scientific contexts

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

    Captions or descriptions of figures or plots

    np_acts-logits-general · gemini-2.0-flash

    scientific and technical figures

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    essex
    -0.37
    campista
    -0.37
     getItemId
    -0.36
    ewel
    -0.35
    mme
    -0.35
    󠁬
    -0.35
     Baillargeon
    -0.35
    moveToFirst
    -0.34
    autocomplete
    -0.34
     συνε
    -0.33
    POSITIVE LOGITS
    StoreMessageInfo
    0.72
    GEBURTSDATUM
    0.68
     betweenstory
    0.66
    Personendaten
    0.61
    évaluateur
    0.61
    fromnode
    0.56
     '\\;'
    0.56
     ModelExpression
    0.53
    IntoConstraints
    0.52
    PerformLayout
    0.51
    Activations Density 0.947%

    No Known Activations