© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 6210
    Prev
    Next
    INDEX
    Explanations

    Random text passages

    np_max-act · gemini-2.0-flash

    This neuron responds to non‐standard tokens such as punctuation marks (especially question marks) and numeric tokens (e.g. years, digits).

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

    tokens that are part of factual informational content (dates, numbers, named places/entities and other salient fact-bearing words).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    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
     oath
    -0.06
    وم
    -0.06
     thresholds
    -0.06
    .advance
    -0.06
    '|
    -0.06
    >New
    -0.06
     Params
    -0.06
     Words
    -0.05
     Insurance
    -0.05
    ["
    -0.05
    POSITIVE LOGITS
     tử
    0.07
    ंबर
    0.06
    เทศ
    0.06
    _NEAR
    0.06
    	ent
    0.06
    irectory
    0.06
    UNIT
    0.06
    اساس
    0.06
    aaaaaaaa
    0.06
    irected
    0.06
    Activations Density 0.775%

    No Known Activations