© 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. 125093
    Prev
    Next
    INDEX
    Explanations

    World

    np_max-act · gemini-2.0-flash

    legal terminology and concepts related to theft and unauthorized use of property.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron consistently activates on instances of the copular verb “is” (and related “be” forms).

    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
    モ
    -0.06
    serrat
    -0.06
     verst
    -0.06
    691
    -0.06
     혹
    -0.06
     head
    -0.06
    373
    -0.06
     throwable
    -0.06
    grund
    -0.06
     _,
    -0.06
    POSITIVE LOGITS
     инвести
    0.07
     ontvang
    0.07
    ीकरण
    0.06
     hüc
    0.06
    scriptions
    0.06
     nucleus
    0.06
     Cobb
    0.06
     homicide
    0.06
    _SWITCH
    0.06
    ampoline
    0.06
    Activations Density 0.359%

    No Known Activations