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. 6687
    Prev
    Next
    INDEX
    Explanations

    references to evidence and proof in relation to claims or assertions

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

    Preceding the word "evidence"

    np_acts-logits-general · gemini-2.0-flash

    providing or denying evidence

    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
    InvalidProtocol
    -0.65
    MigrationBuilder
    -0.65
    EDEFAULT
    -0.62
     inflater
    -0.61
     resourceCulture
    -0.61
    IntoConstraints
    -0.61
     ***!
    -0.60
     surla
    -0.60
     nahilalakip
    -0.60
    serviceWorker
    -0.58
    POSITIVE LOGITS
     evidence
    2.80
     proof
    2.62
    evidence
    2.46
     Evidence
    2.31
    Evidence
    2.31
     EVIDENCE
    2.23
     Proof
    2.13
     PROOF
    2.13
    Proof
    2.12
    proof
    2.10
    Activations Density 1.228%

    No Known Activations