© 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
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 15-RESID-POST-AA
    6. 9372
    Prev
    Next
    INDEX
    Explanations

    Punctuation and "and"

    np_max-act · gemini-2.0-flash

    Tokens that mark clause or sentence boundaries—especially punctuation like commas and periods (and nearby small connective words).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0

    the neuron detects tokens within the assistant's explanatory/informative declarative sentences (assistant-generated explanatory prose).

    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_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     aviation
    -0.07
     wreck
    -0.07
    eph
    -0.06
    error
    -0.06
     relocate
    -0.06
    71
    -0.06
    char
    -0.06
    ोई
    -0.06
    etim
    -0.05
     sph
    -0.05
    POSITIVE LOGITS
    -containing
    0.07
    ,right
    0.07
    "What
    0.06
     справж
    0.06
    llll
    0.06
     lidi
    0.06
     Classification
    0.06
     якому
    0.06
     yö
    0.06
    ximity
    0.06
    Activations Density 0.336%

    No Known Activations