© 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. 7008
    Prev
    Next
    INDEX
    Explanations

    biographical/informational texts

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on numeric tokens and number‐like constructions (e.g. years, scores, “T20,” “limited-overs,” ordinals).

    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
     DUI
    -0.07
    ичного
    -0.07
    eras
    -0.06
    üç
    -0.06
     Malcolm
    -0.06
    _acc
    -0.06
     России
    -0.06
     华
    -0.06
     трьох
    -0.06
    华
    -0.06
    POSITIVE LOGITS
    	super
    0.07
     zvý
    0.06
     lobbying
    0.06
    	throw
    0.06
    .trim
    0.06
     utilizar
    0.06
    }")↵
    0.06
    ΙΑΚ
    0.06
     demise
    0.06
     Cumhurbaşkanı
    0.06
    Activations Density 0.090%

    No Known Activations