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. 14552
    Prev
    Next
    INDEX
    Explanations

    evidence of edits or changes made to documents or statements

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

    time or sequence

    np_acts-logits-general · gemini-2.0-flash

    later, afterwards, or thereafter

    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
    openConnection
    -0.66
    zzleHttp
    -0.60
    encodeWith
    -0.59
    rrggbb
    -0.58
    withIdentifier
    -0.52
     tartalomajánló
    -0.51
    PathParam
    -0.49
    ipedi
    -0.49
    quette
    -0.46
    AlterField
    -0.46
    POSITIVE LOGITS
     later
    3.90
    Later
    3.34
    later
    3.33
     Later
    3.20
     afterwards
    3.12
     afterward
    2.93
     LATER
    2.92
     später
    2.86
     subsequently
    2.86
     subsequent
    2.81
    Activations Density 1.424%

    No Known Activations