© 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. 6497
    Prev
    Next
    INDEX
    Explanations

    punctuation and formatting symbols

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

    Something before a method or function call

    np_acts-logits-general · gemini-2.0-flash

    programming assertions and display

    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
    oredCriteria
    -0.82
     itſelf
    -0.80
     Shakspeare
    -0.79
     صوتيه
    -0.76
     feroit
    -0.75
    \{\\
    -0.75
     industriale
    -0.74
    RTEX
    -0.72
     Monfieur
    -0.72
     Jefus
    -0.72
    POSITIVE LOGITS
     fact
    0.63
     co
    0.53
    nah
    0.49
     log
    0.49
     Ce
    0.48
    fact
    0.48
     focus
    0.47
     mess
    0.46
     sleep
    0.46
    AccessorTable
    0.45
    Activations Density 0.094%

    No Known Activations