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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 18-GEMMASCOPE-RES-16K
    6. 1394
    Prev
    Next
    INDEX
    Explanations

    positive and negative sentiments in product reviews

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

    Punctuation followed by discourse markers

    np_acts-logits-general · gemini-2.0-flash

    sentence end punctuation

    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
    Compartir
    -0.62
    apter
    -0.49
    Executives
    -0.48
    keyColumn
    -0.46
     mitte
    -0.46
    styleType
    -0.43
    odotus
    -0.43
    nachron
    -0.42
    bigoplus
    -0.42
    tact
    -0.42
    POSITIVE LOGITS
     got
    1.02
     gotta
    0.97
     still
    0.89
     Took
    0.89
     Got
    0.87
     gonna
    0.87
     Been
    0.86
    Gonna
    0.84
     Cannot
    0.84
     never
    0.83
    Activations Density 0.389%

    No Known Activations