© 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
    2. Gemma-2-2B
    3. 25-GEMMASCOPE-TRANSCODER-16K
    4. 3930
    Prev
    Next
    INDEX
    Explanations

    words indicating condition, invention, and other somewhat random words, with a strong bias toward the word "will"

    oai_token-act-pair · gemini-2.0-flash

    variety of text snippets

    np_max-act-logits · gemini-2.0-flash

    auxiliary verbs

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.25.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
     choose
    -0.71
     understanding
    -0.69
     understand
    -0.68
     choosing
    -0.67
     know
    -0.65
     chooses
    -0.65
     understands
    -0.65
     preferring
    -0.63
     spend
    -0.61
     enjoying
    -0.61
    POSITIVE LOGITS
     calendriers
    0.75
    KURZBESCHREIBUNG
    0.71
     lenker
    0.71
    LookAnd
    0.69
     ligiloj
    0.67
     חיצוניים
    0.64
     EconPapers
    0.63
    حياتها
    0.63
    లాలు
    0.61
     Hurts
    0.59
    Activations Density 12.797%

    No Known Activations