© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 2868
    Prev
    Next
    INDEX
    Explanations

    special formatting characters or symbols used in an encoded context

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

    ending in "self" or containing math symbols

    np_acts-logits-general · gemini-2.0-flash

    jurisdictional claims

    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
     R
    -0.57
     "
    -0.56
     
    -0.54
     K
    -0.54
     (
    -0.54
     T
    -0.53
     V
    -0.53
     Pat
    -0.52
    ,
    -0.52
     E
    -0.52
    POSITIVE LOGITS
     myſelf
    1.11
     itſelf
    1.02
     للاسماء
    1.01
     themſelves
    1.00
     CreateTagHelper
    0.97
     Theſe
    0.96
     propOrder
    0.95
    ſelf
    0.92
     Efq
    0.91
     auffi
    0.90
    Activations Density 0.001%

    No Known Activations