© 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. 31054
    Prev
    Next
    INDEX
    Explanations

    Continuing a thought

    np_max-act · gemini-2.0-flash

    The neuron is selectively firing on sentence‐ or clause‐boundary punctuation (periods, commas) and the capitalized tokens that tend to follow them, effectively detecting sentence breaks.

    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
     AQ
    -0.06
    612
    -0.06
     Thomson
    -0.06
    З
    -0.06
    itag
    -0.06
    Seattle
    -0.06
     Matching
    -0.06
    ERICA
    -0.06
     declines
    -0.06
     کلاس
    -0.06
    POSITIVE LOGITS
    -of
    0.07
    ternet
    0.07
    .command
    0.07
    -sw
    0.06
    legalArgumentException
    0.06
     geral
    0.06
    切り
    0.06
     unsupported
    0.06
     св
    0.06
    _TD
    0.06
    Activations Density 0.117%

    No Known Activations