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. 3890
    Prev
    Next
    INDEX
    Explanations

    elements related to game mechanics and dice operations

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

    Probability calculations of picking letters

    np_acts-logits-general · gemini-2.0-flash

    rolling dice and coin flips

    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
    parametrize
    -0.51
     vPvB
    -0.47
    ylem
    -0.47
    綜合報導
    -0.46
    contentValues
    -0.44
    ροφο
    -0.44
    quíni
    -0.44
    גרפיה
    -0.44
    lossians
    -0.43
    ArgumentParser
    -0.42
    POSITIVE LOGITS
     roll
    2.21
     rolls
    2.07
     rolled
    2.00
     Roll
    1.91
     rolling
    1.90
    Roll
    1.82
    roll
    1.80
     ROLL
    1.79
     Rolling
    1.70
    ROLL
    1.63
    Activations Density 0.171%

    No Known Activations