© 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. 6159
    Prev
    Next
    INDEX
    Explanations

    lli

    np_max-act · gemini-2.0-flash

    This neuron detects early-clause evaluative or hedging constructions, especially phrases starting with “It is …” (e.g. “It is important to…,” “It is difficult to…”).

    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
    .
    -0.07
     ceilings
    -0.07
     человек
    -0.07
    े।
    -0.06
    .
    -0.06
    icked
    -0.06
    och
    -0.06
     kuv
    -0.06
    μαν
    -0.06
     Barrel
    -0.06
    POSITIVE LOGITS
    ={{
    0.07
    (clone
    0.07
     detached
    0.06
     getenv
    0.06
     uniq
    0.06
    _SANITIZE
    0.06
     trữ
    0.06
    истра
    0.06
     butterknife
    0.06
    /license
    0.06
    Activations Density 3.351%

    No Known Activations