© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 28-GEMMASCOPE-RES-16K
    11. 4307
    Prev
    Next
    INDEX
    Explanations

    the word "which" in various forms and contexts

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

    "which" followed by a clause

    np_acts-logits-general · gemini-2.0-flash

    which or where introducing a clause

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 28-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_28/width_16k/average_l0_119
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.28.hook_resid_post
    Hook Layer
    28
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     المعيارى
    -0.65
     pinulongan
    -0.54
     mergeFrom
    -0.46
     Lightboxes
    -0.45
    btnBack
    -0.45
     rhestr
    -0.44
     appelez
    -0.44
    Ằ
    -0.44
    évaluateur
    -0.43
    pero
    -0.42
    POSITIVE LOGITS
     whereof
    0.61
     we
    0.61
     they
    0.57
     duquel
    0.56
     you
    0.54
    soever
    0.54
     everybody
    0.47
    where
    0.44
    ObjectModel
    0.43
    wherein
    0.43
    Activations Density 0.164%

    No Known Activations