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. 20-GEMMASCOPE-RES-16K
    6. 2653
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate exceptions or deviations from general rules or norms

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

    "exception" or "exceptions"

    np_acts-logits-general · gemini-2.0-flash

    "exception" or "exceptions"

    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
     besk
    -0.40
     kí
    -0.39
     piú
    -0.37
    quetas
    -0.37
     reprend
    -0.37
    LabelTagHelper
    -0.37
    oya
    -0.37
    seamnă
    -0.36
    構
    -0.35
    ActionCreators
    -0.35
    POSITIVE LOGITS
     exceptions
    3.16
     exception
    3.04
     Exceptions
    2.61
     excep
    2.28
    exception
    2.27
     Exception
    2.27
    Exceptions
    2.26
    exceptions
    2.23
     excepción
    2.10
     exemption
    2.09
    Activations Density 0.534%

    No Known Activations