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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 1953
    Prev
    Next
    INDEX
    Explanations

    technical descriptions of systems involving connections and infrastructures

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

    cables, lines, and connections

    np_acts-logits-general · gemini-2.0-flash

    electrical transmission lines and cables

    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
    DataAnnotations
    -0.52
    æv
    -0.45
    tidase
    -0.43
    OnLoad
    -0.43
    fromnode
    -0.42
    rikes
    -0.42
     furnaces
    -0.42
    DataAdapter
    -0.41
    ikkert
    -0.41
     nuc
    -0.41
    POSITIVE LOGITS
     cable
    1.32
     cables
    1.28
     cabling
    1.20
     wires
    1.18
     wire
    1.15
     Cables
    1.05
     Cable
    1.03
     wiring
    1.03
    cable
    1.01
    cables
    0.99
    Activations Density 0.473%

    No Known Activations