© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5466
    Prev
    Next
    INDEX
    Explanations

    purpose related words

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

    The neuron lights up on the chief “head” noun in descriptive or promotional phrases—i.e. the key product‐ or concept‐type noun (finish, solution, distribution, cities, styles, sound, relationships, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    かもしれませんが
    0.89
    ;//
    0.83
    ですし
    0.82
     কেননা
    0.80
    はもちろん
    0.79
    because
    0.75
     inasmuch
    0.75
    이니까
    0.74
     অথবা
    0.73
    ];//
    0.73
    POSITIVE LOGITS
     turnkey
    0.89
     हजारों
    0.87
     vele
    0.85
     மிகப்பெரிய
    0.85
     rapidement
    0.80
     brandy
    0.79
    完整的
    0.76
     பெரும்
    0.73
     tesla
    0.73
     sonrası
    0.72
    Activations Density 0.417%

    No Known Activations