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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Gemma-2-2B
    3. 16-GEMMASCOPE-TRANSCODER-16K
    4. 1095
    Prev
    Next
    INDEX
    Explanations

    review

    np_max-act-logits · gemini-2.0-flash

    things that were inaccurate as part of reviews, rebuttals, or official statements

    oai_token-act-pair · gemini-2.0-flash

    Something related to film reviews, statements, batteries, reporting and claims. It also activates for almost any number ending in 4, 5, 6, 7, 8, 9, or 0. It also activates for some punctuation/special characters. This neuron activates for numerals

    np_token-act-pair-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_16/width_16k/average_l0_10
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.16.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
     disambiguazione
    -0.72
     escoger
    -0.70
     though
    -0.70
     Though
    -0.69
     incluyendo
    -0.68
    Though
    -0.66
    تقاوى
    -0.63
     tới
    -0.62
    новништво
    -0.60
     πως
    -0.59
    POSITIVE LOGITS
    NUMX
    1.25
     XNUMX
    1.20
     ​​
    1.00
     .;
    0.80
    ̵
    0.72
     ™
    0.71
    ​​
    0.69
     .:
    0.67
     ?!
    0.64
     myſelf
    0.63
    Activations Density 5.876%

    No Known Activations