© 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
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 92033
    Prev
    Next
    INDEX
    Explanations

    .

    np_max-act · gemini-2.0-flash

    The neuron lights up on sentence-level discourse markers and key structural words—things like “In,” “as,” “crucial,” “therefore,” “however,” and the closing period—i.e. transition cues that organize and emphasize the flow of an academic argument.

    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
     evening
    -0.07
     caused
    -0.06
     eternal
    -0.06
     cause
    -0.06
     parsing
    -0.06
     Turk
    -0.06
     Campos
    -0.06
     Wilhelm
    -0.06
    _loop
    -0.06
     little
    -0.06
    POSITIVE LOGITS
    ameleon
    0.07
    \">"
    0.07
    Cancelar
    0.07
     статті
    0.07
    systems
    0.07
    bett
    0.06
    communic
    0.06
     dentro
    0.06
    にする
    0.06
    svp
    0.06
    Activations Density 0.116%

    No Known Activations