© 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. 0-GEMMASCOPE-TRANSCODER-16K
    7. 6394
    Prev
    Next
    INDEX
    Explanations

    the string "IP" or "IR" within longer all-caps strings

    oai_token-act-pair · gemini-2.0-flash

    Code/Data structures

    np_max-act · gemini-2.0-flash

    kernel/network constants

    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_0/width_16k/average_l0_76
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.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
     IC
    -1.17
     IM
    -1.15
     IB
    -1.13
     IP
    -1.08
     IR
    -1.03
     IA
    -1.01
     IK
    -0.98
    Ｉ
    -0.97
    IC
    -0.95
     IS
    -0.94
    POSITIVE LOGITS
     Lipa
    0.65
    TargetException
    0.65
    if
    0.63
    iii
    0.63
    ii
    0.63
    itm
    0.61
    ieR
    0.60
    MessageTagHelper
    0.60
    is
    0.59
    AntiForgeryToken
    0.59
    Activations Density 1.276%

    No Known Activations