© 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. 5639
    Prev
    Next
    INDEX
    Explanations

    words and short dependent clauses that express reasoning and justification

    oai_token-act-pair · gemini-2.0-flash

    specific fields or categories

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

    hobbies and discovery

    np_max-act-logits · gemini-2.0-flash

    hobbies and professions Method used: 3 — because top logits list activities/fields; explanation: hobbies and professions

    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_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
     compagni
    -0.60
     découver
    -0.59
     restantes
    -0.58
     fidèles
    -0.56
     copertura
    -0.55
     trasmissione
    -0.54
     sagesse
    -0.53
     soyez
    -0.53
     camada
    -0.53
     plupart
    -0.53
    POSITIVE LOGITS
    GenerationType
    0.85
     aviation
    0.83
    Autoritní
    0.82
     yoga
    0.80
     astronomy
    0.79
     dentistry
    0.75
     skateboarding
    0.74
    orescence
    0.73
     autorytatywna
    0.73
     Italijanski
    0.73
    Activations Density 6.669%

    No Known Activations