© 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. 41873
    Prev
    Next
    INDEX
    Explanations

    how something is done

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

    The neuron detects explanatory or procedural cues in text—most strongly activating on occurrences of the interrogative “how” (e.g. “how to…,” “how these…,” 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
     philippe
    -1.67
     lopez
    -1.60
    
    -1.59
     andré
    -1.58
     felipe
    -1.51
     Tél
    -1.49
    骓
    -1.48
     disputa
    -1.46
     ricardo
    -1.45
    ötä
    -1.45
    POSITIVE LOGITS
     they
    2.58
     that
    1.98
    Each
    1.88
     verschiedenen
    1.77
     різних
    1.73
     diferentes
    1.72
     einzelnen
    1.70
    いくつかの
    1.69
     distintas
    1.68
     jeweiligen
    1.63
    Activations Density 0.019%

    No Known Activations