© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 19742
    Prev
    Next
    INDEX
    Explanations

    Conn, eslint, or date contexts

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on any token containing the substring “conn” (e.g. “Conn,” “socket.Conn,” “invalid Conn type,” “Newtown, Conn,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     and
    -2.66
    翃
    -2.64
    迓
    -2.56
    ે
    -2.44
    и
    -2.33
    <<<<<<<
    -2.30
    л
    -2.28
    ма
    -2.23
    ion
    -2.17
    全新的
    -2.16
    POSITIVE LOGITS
    橚
    3.59
    釭
    3.48
    鉿
    3.08
    咴
    2.78
    𓆟
    2.58
    毉
    2.53
    卝
    2.41
    ﹍﹍
    2.38
    VERY
    2.38
     ecuator
    2.36
    Activations Density 0.001%

    No Known Activations