Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 12726
    Prev
    Next
    INDEX
    Explanations

    intensifiers and adverbs that amplify expression

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

    "So" followed by an adverb or pronoun

    np_acts-logits-general · gemini-2.0-flash

    so far and such as

    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
    agisse
    -0.54
    żeli
    -0.53
    SearchParams
    -0.52
    ittää
    -0.50
     ['./
    -0.49
     lagar
    -0.49
    miri
    -0.49
    [__
    -0.48
    Története
    -0.48
    scriptcase
    -0.48
    POSITIVE LOGITS
     much
    0.95
    othed
    0.83
     far
    0.76
     darn
    0.70
    much
    0.68
     damn
    0.68
    ckeye
    0.67
    UnknownFieldSet
    0.66
    Physician
    0.66
    Much
    0.65
    Activations Density 0.095%

    No Known Activations