© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 110281
    Prev
    Next
    INDEX
    Explanations

    Nonsense

    np_max-act · gemini-2.0-flash

    The main thing this neuron does is detect special internal/control tokens (e.g. header or metadata markers) in the text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    requests for or generation of deliberately nonsensical, absurd, or intentionally incorrect responses.

    oai_token-act-pair · gpt-5Triggered 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
    •Misaligned Persona
    •my list

    No Comments

    Negative Logits
    Once
    -0.07
     Nearly
    -0.06
    аними
    -0.06
     Zoe
    -0.06
     Namespace
    -0.06
     Pakistani
    -0.06
     smoothly
    -0.06
    _rates
    -0.06
    TextLabel
    -0.06
    .Modules
    -0.06
    POSITIVE LOGITS
    Но
    0.07
    !(:
    0.07
     strtol
    0.06
    (fig
    0.06
    ();?>
    0.06
    恶
    0.06
    [dir
    0.06
    ovaná
    0.06
     aload
    0.06
    ults
    0.06
    Activations Density 0.055%

    No Known Activations