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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 21-GEMMASCOPE-TRANSCODER-16K
    7. 5554
    Prev
    Next
    INDEX
    Explanations

    words often appearing in academic writing, such as prepositions, conjunctions, and words expressing an author's intent

    oai_token-act-pair · gemini-2.0-flash

    present participles related to actions

    np_token-act-pair-logits · gpt-4o-mini

    of

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     to
    -0.57
     questão
    -0.40
     revés
    -0.39
     hely
    -0.38
     possibilidade
    -0.38
     excepción
    -0.37
     fasse
    -0.37
    h
    -0.36
    to
    -0.36
     disposição
    -0.36
    POSITIVE LOGITS
     creating
    1.51
     providing
    1.51
     bringing
    1.47
     delivering
    1.46
     selecting
    1.44
     introducing
    1.43
     displaying
    1.43
     producing
    1.43
     releasing
    1.41
     discovering
    1.39
    Activations Density 11.281%

    No Known Activations