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. 22-GEMMASCOPE-RES-16K
    6. 10282
    Prev
    Next
    INDEX
    Explanations

    instructions and guidelines related to the safe use of a substance

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

    "Meta" or "Met" prefixes/words

    np_acts-logits-general · gemini-2.0-flash

    meta and met words

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Sai
    -0.66
     Barbour
    -0.64
    󠁢
    -0.63
    ussis
    -0.58
    Sai
    -0.57
    esgue
    -0.56
     Cray
    -0.56
     Autorizaciones
    -0.55
     Dall
    -0.55
     Bav
    -0.55
    POSITIVE LOGITS
     MET
    1.96
     Meta
    1.90
    Meta
    1.78
     meta
    1.76
    MET
    1.76
     Met
    1.72
     META
    1.71
    meta
    1.65
    Met
    1.60
     metadata
    1.57
    Activations Density 0.340%

    No Known Activations