© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 6820
    Prev
    Next
    INDEX
    Explanations

    think

    np_max-act · gemini-2.0-flash

    The neuron fires on hedging or opinion phrases—especially first-person verbs like “think,” “suppose,” “guess,” etc., that signal the author’s subjective or uncertain statements.

    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
    internet
    -0.07
    andre
    -0.06
    chner
    -0.06
    获取
    -0.06
    ством
    -0.06
    орту
    -0.06
     процесу
    -0.06
    FORM
    -0.06
    Pakistan
    -0.06
    Dat
    -0.06
    POSITIVE LOGITS
    .columnHeader
    0.07
     plag
    0.07
     Tribal
    0.06
     Input
    0.06
     subsidy
    0.06
     attire
    0.06
    0.06
     surfaced
    0.06
    \',
    0.06
     Classes
    0.06
    Activations Density 0.018%

    No Known Activations