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. 10459
    Prev
    Next
    INDEX
    Explanations

    instances of formal or technical references in text, particularly those related to methods or actions taken in systems or processes

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

    "Bottom line" followed by a statement

    np_acts-logits-general · gemini-2.0-flash

    bottom line

    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
     ujednoznacz
    -0.71
    DockStyle
    -0.69
    uxxxx
    -0.61
    TagHelper
    -0.61
     intptr
    -0.61
    WriteTagHelper
    -0.59
    autoreleasepool
    -0.57
    tvguidetime
    -0.57
     nahilalakip
    -0.56
    談社
    -0.54
    POSITIVE LOGITS
     line
    3.31
     lines
    2.95
    line
    2.86
     Line
    2.79
    Line
    2.63
    lines
    2.57
     LINE
    2.55
     Lines
    2.54
    LINE
    2.48
    Lines
    2.35
    Activations Density 0.648%

    No Known Activations