© 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. 8204
    Prev
    Next
    INDEX
    Explanations

    prepositions and post-modifiers

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

    The neuron fires on passages where someone is describing their own investment portfolio or giving personal‐finance details (e.g. “My portfolio is entirely in funds…,” links/access to research, “you might think I am losing by having high expenses,” etc.). In other words, it spotlights first‐person personal‐finance commentary and portfolio descriptions.

    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
     accomplishments
    0.86
     oftentimes
    0.80
    Lorsque
    0.79
    Toward
    0.78
    Zum
    0.77
    Rosen
    0.74
    ]^{
    0.73
     splendor
    0.73
    具有
    0.73
    Beginning
    0.73
    POSITIVE LOGITS
     dodgy
    1.45
     motorway
    1.19
     bloke
    1.13
     postcode
    1.13
     saddo
    1.12
     всё
    1.11
     hugely
    1.11
     carers
    1.10
     maths
    1.08
     underpinned
    1.07
    Activations Density 0.052%

    No Known Activations