© 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. 128737
    Prev
    Next
    INDEX
    Explanations

    format or template

    np_max-act · gemini-2.0-flash

    This neuron detects the explicit “format:” instruction (i.e. the word “format” and its following colon) used to introduce the required answer template.

    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
    штов
    -0.07
     sire
    -0.07
     CAL
    -0.07
     timeZone
    -0.07
     Goal
    -0.07
    _IGNORE
    -0.07
    Science
    -0.06
    소개
    -0.06
     spar
    -0.06
    BoxLayout
    -0.06
    POSITIVE LOGITS
    оти
    0.06
    ες
    0.06
     poorest
    0.06
     asymmetric
    0.06
     Modified
    0.06
    larda
    0.06
     contexto
    0.06
     Dort
    0.05
     возв
    0.05
     Fort
    0.05
    Activations Density 0.008%

    No Known Activations