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. 2673
    Prev
    Next
    INDEX
    Explanations

    words related to specific entities or identifiers, particularly proper nouns

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

    Strings followed by suffixes like "ue", "uk", or "arsk"

    np_acts-logits-general · gemini-2.0-flash

    ts Ts ts ts

    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
     exécut
    -0.56
     CHAPITRE
    -0.51
     quelcon
    -0.51
     Bower
    -0.50
     Wurzel
    -0.50
     speranza
    -0.50
     exemplaire
    -0.48
     campion
    -0.48
     fär
    -0.48
     exemplaires
    -0.48
    POSITIVE LOGITS
    IBOutlet
    0.81
    دانشنامهٔ
    0.74
     kasarigan
    0.72
     TS
    0.72
     ts
    0.71
     Ts
    0.70
    Хьажоргаш
    0.69
    KommentareTeilen
    0.68
     hObject
    0.66
    hirts
    0.63
    Activations Density 0.105%

    No Known Activations