© 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. 8832
    Prev
    Next
    INDEX
    Explanations

    The

    np_max-act · gemini-2.0-flash

    This neuron detects sentence-initial discourse markers and common transition words (e.g. “In,” “And,” “After,” “Additionally,” “So,” “About,” “The”) that signal the start of a new statement.

    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
    avourite
    -0.07
     Construction
    -0.07
    ish
    -0.07
    .amazon
    -0.07
     counterpart
    -0.07
    .Then
    -0.06
    /top
    -0.06
    oftware
    -0.06
     cousin
    -0.06
     bipartisan
    -0.06
    POSITIVE LOGITS
     temperament
    0.06
     emacs
    0.06
     jclass
    0.06
     medya
    0.06
     M
    0.06
     L
    0.06
     Она
    0.06
     pornô
    0.06
     cravings
    0.06
    Dat
    0.06
    Activations Density 0.011%

    No Known Activations