© 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. 22-GEMMASCOPE-RES-16K
    6. 5359
    Prev
    Next
    INDEX
    Explanations

    structured code patterns in a programming context

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

    ending in "Listener"

    np_acts-logits-general · gemini-2.0-flash

    listener and error handling

    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
    $_['
    -0.52
    protoimpl
    -0.51
     Marke
    -0.48
    Marke
    -0.46
    ]
    -0.44
     bern
    -0.43
    ]\
    -0.43
    ponses
    -0.43
    =?";
    -0.43
     Horn
    -0.43
    POSITIVE LOGITS
    ergies
    0.69
    FailureListener
    0.66
    icoot
    0.65
    cotch
    0.64
    ExtendWith
    0.64
    ELTS
    0.62
     Taktlose
    0.59
    InvalidProtocol
    0.59
    ErrorListener
    0.59
     Emits
    0.58
    Activations Density 0.051%

    No Known Activations