© 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. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 124957
    Prev
    Next
    INDEX
    Explanations

    Spoken/written language

    np_max-act · gemini-2.0-flash

    The neuron fires on tokens involved in meta-linguistic discussion—words like “language,” “words,” “impact,” and “mindful”—i.e. when the text is talking about word choice, respectful wording, or the effects of language.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    language that promotes respectful and affirming communication.

    oai_token-act-pair · gpt-4o-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
    ctrine
    -0.07
     medieval
    -0.06
    ่าท
    -0.06
    mousedown
    -0.06
    Ğİ
    -0.06
     ServletException
    -0.06
     Recycling
    -0.06
    ReadOnly
    -0.06
    _fre
    -0.06
    .Download
    -0.06
    POSITIVE LOGITS
    Trail
    0.07
     Pixar
    0.07
     Ill
    0.07
     аром
    0.06
    рой
    0.06
    Link
    0.06
     Eugene
    0.06
    zahl
    0.06
    										↵
    0.06
     Courts
    0.06
    Activations Density 0.020%

    No Known Activations