© 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. 5122
    Prev
    Next
    INDEX
    Explanations

    references to construction and architectural features

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

    Units of measure or containers

    np_acts-logits-general · gemini-2.0-flash

    parts of objects

    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
     linkovi
    -0.70
    culable
    -0.59
    écri
    -0.58
     Serikat
    -0.57
     playera
    -0.56
     Generalized
    -0.56
     eletrônico
    -0.55
     becauſe
    -0.55
    swire
    -0.54
     adentro
    -0.54
    POSITIVE LOGITS
    Kg
    0.56
     £
    0.56
     &___
    0.56
     skimmed
    0.53
     Kg
    0.52
     hefyd
    0.50
     UIAlert
    0.49
    ChildIndex
    0.49
    Fitted
    0.49
    Fixing
    0.49
    Activations Density 0.158%

    No Known Activations