© 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. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 41028
    Prev
    Next
    INDEX
    Explanations

    fish

    np_max-act · gemini-2.0-flash

    This neuron fires on whole “content” words (especially standalone nouns or other meaningful terms) and stays off on subword fragments and common function words.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     LOCK
    -0.08
    œ
    -0.07
     ﾚ
    -0.07
     decentralized
    -0.06
    icopter
    -0.06
     Entered
    -0.06
    なかった
    -0.06
    叫
    -0.06
     RED
    -0.05
     commitment
    -0.05
    POSITIVE LOGITS
    angelog
    0.07
    ultiple
    0.07
     tamb
    0.06
     функ
    0.06
    venience
    0.06
     intuit
    0.06
    ,W
    0.06
    \xb
    0.06
    .setState
    0.06
    Sl
    0.06
    Activations Density 0.095%

    No Known Activations