Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 7520
    Prev
    Next
    INDEX
    Explanations

    references to legal provisions and their implications within a document

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

    Preceding relative pronouns ("of which", "to whom")

    np_acts-logits-general · gemini-2.0-flash

    relative clauses introduced by which or whom

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ządu
    -0.53
    AutoField
    -0.48
    IsPostBack
    -0.47
    ValueGenerated
    -0.46
    arschijnlijk
    -0.46
    umą
    -0.45
    CppCodeGen
    -0.44
     actionTypes
    -0.43
    insertBefore
    -0.43
    :///
    -0.43
    POSITIVE LOGITS
     which
    2.84
    which
    2.19
     Which
    2.10
     WHICH
    2.10
    Which
    1.99
     whom
    1.64
    hich
    1.51
     laquelle
    1.48
     lequel
    1.48
     whereof
    1.47
    Activations Density 0.320%

    No Known Activations