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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 18-GEMMASCOPE-RES-16K
    6. 745
    Prev
    Next
    INDEX
    Explanations

    statements and evidence related to claims of fraud or lack of substantiation

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

    Evidence or convincing evidence

    np_acts-logits-general · gemini-2.0-flash

    evidence

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    
    -1.03
    IntoConstraints
    -0.81
    ⓧ
    -0.77
     inflater
    -0.64
     nahilalakip
    -0.63
    ContentAsync
    -0.62
    MigrationBuilder
    -0.61
    djangoproject
    -0.60
     ***!
    -0.59
    stylers
    -0.59
    POSITIVE LOGITS
     evidence
    2.25
     proof
    2.12
    evidence
    2.00
    Evidence
    1.95
     Evidence
    1.92
     EVIDENCE
    1.85
     proofs
    1.77
     Proof
    1.73
    Proof
    1.73
    proof
    1.71
    Activations Density 0.964%

    No Known Activations