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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9453
    Prev
    Next
    INDEX
    Explanations

    numerical data related to complex mathematical concepts and representations

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

    Mathematical tuples or cycles

    np_acts-logits-general · gemini-2.0-flash

    tuples, cycles, dimensions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    LError
    -0.49
    tahui
    -0.45
    Solubility
    -0.43
    FindBy
    -0.43
     alguna
    -0.41
     kader
    -0.41
     letto
    -0.40
    География
    -0.40
    assembling
    -0.39
    Famille
    -0.39
    POSITIVE LOGITS
    tagHelperRunner
    0.93
     تضيفلها
    0.84
     AssemblyCulture
    0.77
     dimensional
    0.73
     disambiguazione
    0.69
    LookAnd
    0.65
    addPreferredGap
    0.64
    ArrowToggle
    0.63
    addContainerGap
    0.63
     defaultstate
    0.61
    Activations Density 1.309%

    No Known Activations