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. 20-GEMMASCOPE-RES-16K
    6. 1469
    Prev
    Next
    INDEX
    Explanations

    specific words indicating identity, cultural references, and socioeconomic contrasts

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

    Two letter prefixes

    np_acts-logits-general · gemini-2.0-flash

    crushing, scariest, bionic, raunchy, pacing

    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
    GEBURTSDATUM
    -0.97
     becauſe
    -0.77
     ſtate
    -0.74
     lenker
    -0.74
    tonode
    -0.72
     theſe
    -0.71
    Portail
    -0.71
     alſo
    -0.70
     ſeveral
    -0.70
     purpoſe
    -0.69
    POSITIVE LOGITS
    entibus
    0.53
     تكبرها
    0.52
     šķ
    0.46
     Zy
    0.45
    eſ
    0.45
    çado
    0.45
     ulei
    0.45
     navigazione
    0.43
    ()");
    0.42
     comprimento
    0.42
    Activations Density 0.707%

    No Known Activations