© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 74624
    Prev
    Next
    INDEX
    Explanations

    describing intended purpose

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

    The neuron detects boilerplate purpose statements, especially the “intended as a … to …” phrasing that signals the text’s stated intent or function.

    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
     want
    -1.84
     decide
    -1.52
                
    -1.46
     elevate
    -1.41
     amplia
    -1.41
     respective
    -1.36
     then
    -1.36
     ensure
    -1.31
     choose
    -1.30
     perceive
    -1.27
    POSITIVE LOGITS
    .”
    2.00
    }=\
    1.46
    ᯅ
    1.41
     and
    1.38
    ')
    1.38
    ');
    1.35
    lässt
    1.34
     >::
    1.34
    τές
    1.33
    !!!!!
    1.33
    Activations Density 0.032%

    No Known Activations