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. 8215
    Prev
    Next
    INDEX
    Explanations

    updates and ratings related to software releases and user experiences

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

    followed by punctuation marks

    np_acts-logits-general · gemini-2.0-flash

    status, events, and qualifiers

    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
    DrawerToggle
    -0.64
    Autoritní
    -0.58
    
    -0.56
    utsal
    -0.55
    ')}}">
    -0.54
     NSError
    -0.54
     nearest
    -0.54
    /#{
    -0.54
    awaiter
    -0.53
    ')}}"
    -0.53
    POSITIVE LOGITS
    存于互联网档案馆
    0.63
    ljiv
    0.55
    且
    0.55
    地
    0.55
     AssemblyCulture
    0.54
    Yet
    0.52
    وردار
    0.52
     pula
    0.51
     namun
    0.50
    Doch
    0.49
    Activations Density 0.030%

    No Known Activations