© 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. Gemma-2-2B
    5. 21-GEMMASCOPE-TRANSCODER-16K
    7. 4787
    Prev
    Next
    INDEX
    Explanations

    passages related to personal feelings, like feeling obligated to text someone back, feeling hungry, or hating Mondays.

    oai_token-act-pair · gemini-2.0-flash

    contractions or possessives

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

    contractions with apostrophes

    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
     olsun
    -0.47
     disait
    -0.46
    stdin
    -0.46
    NOPQRST
    -0.46
     jäsen
    -0.45
    aurait
    -0.44
     kanssa
    -0.44
    iança
    -0.43
     arvio
    -0.43
    AnchorStyles
    -0.43
    POSITIVE LOGITS
     hes
    1.91
     shes
    1.80
     Theres
    1.77
    Theres
    1.75
     its
    1.74
     theyre
    1.74
     youre
    1.71
     Thats
    1.69
     Ive
    1.68
     Im
    1.67
    Activations Density 1.665%

    No Known Activations