© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 128613
    Prev
    Next
    INDEX
    Explanations

    starting words of phrases

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

    The neuron is detecting future‐tense auxiliaries – it lights up on English “will” and on Malay/Indonesian “akan” (and even the same letter sequence when it appears in other tokens).

    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
     heath
    -2.45
     foggy
    -2.41
     covet
    -2.39
    းက
    -2.30
    ).
    -2.23
    ",
    -2.20
    \
    -2.19
    ៚
    -2.19
     compulsive
    -2.13
    ve
    -2.11
    POSITIVE LOGITS
     be
    3.13
     tetapi
    2.61
    很喜欢
    2.47
     )
    2.38
     ');
    2.28
    Ꮉ
    2.23
     he
    2.20
     Our
    2.09
    ))
    2.09
     );
    2.06
    Activations Density 0.000%

    No Known Activations