© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 14152
    Prev
    Next
    INDEX
    Explanations

    hard to accomplish things

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

    The neuron detects phrasing that expresses difficulty or challenge (e.g. “hard to…,” “difficult to…,” “hardest to accomplish in…,” 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
     ï
    -1.48
     ‚
    -1.46
     interesses
    -1.40
    禟
    -1.39
     disminución
    -1.37
    を選択
    -1.37
     renfer
    -1.35
    部分が
    -1.34
     różnych
    -1.34
     participer
    -1.34
    POSITIVE LOGITS
     in
    1.90
     people
    1.77
     as
    1.59
     would
    1.38
     seems
    1.30
     and
    1.28
     a
    1.25
    子
    1.22
    VolleyError
    1.21
    mathrm
    1.20
    Activations Density 0.085%

    No Known Activations