© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 7878
    Prev
    Next
    INDEX
    Explanations

    instructions

    np_max-act · gemini-2.0-flash

    This neuron fires on imperative action verbs that begin steps in instructional or how-to lists.

    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
    elda
    -0.07
    .getValue
    -0.06
    .DataType
    -0.06
    기타
    -0.06
    '>↵
    -0.06
    TimeStamp
    -0.06
    imals
    -0.06
    onta
    -0.06
     stance
    -0.06
    .While
    -0.06
    POSITIVE LOGITS
     توان
    0.06
    _CLEAR
    0.06
     iliş
    0.06
     канди
    0.06
    duğ
    0.06
    bud
    0.06
    бин
    0.06
     Conj
    0.06
     klas
    0.06
     действ
    0.06
    Activations Density 0.053%

    No Known Activations