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. 25-GEMMASCOPE-RES-16K
    6. 6064
    Prev
    Next
    INDEX
    Explanations

    the presence of the word "American" in various contexts

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

    American

    np_max-act-logits · gpt-4.1-mini-2025-04-14Triggered by @arientigiacomo2003

    The token consistently appears as an initial capitalized adjective referring to organizations, associations, institutions, programs, or descriptors affiliated with the United States. It commonly precedes formal group names, titles, or proper nouns indicating national identity or origin. This pattern occurs across diverse domains including healthcare, education, scientific societies, historical and cultural references, and professional bodies.

    eleuther_acts_top20 · gpt-4.1-mini-2025-04-14Triggered by @arientigiacomo2003
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 25-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_25/width_16k/average_l0_116
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     виправивши
    -0.57
    })();
    
    -0.54
    cleros
    -0.53
    EndInit
    -0.53
     matter
    -0.53
    Tikang
    -0.52
     GenerationType
    -0.51
    cales
    -0.49
    Iné
    -0.49
     HttpServlet
    -0.49
    POSITIVE LOGITS
     Samoa
    0.74
    usermodel
    0.59
    Samoa
    0.56
    stateProvider
    0.55
    RegressionTest
    0.54
    зулта
    0.53
     Idol
    0.53
    ism
    0.52
     Serikat
    0.52
     Legion
    0.52
    Activations Density 0.176%

    No Known Activations