© 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
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 12177
    Prev
    Next
    INDEX
    Explanations

    first-person pronoun

    np_max-act · gemini-2.0-flash

    inconsistencies and missing values in data tables.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron fires on inline numeric probability/weight tokens—i.e. decimal numbers (like 0.307861328125, 2.04296875, etc.) that encode model‐internal scores or metadata.

    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
    ussels
    -0.06
     jus
    -0.06
    鐵
    -0.06
    iyim
    -0.06
    	rep
    -0.06
     留
    -0.06
            		
    -0.06
     открыт
    -0.06
    .are
    -0.06
    ..."↵
    -0.06
    POSITIVE LOGITS
     "./
    0.07
     toxin
    0.07
     moth
    0.07
     Serialized
    0.06
     archive
    0.06
     Cal
    0.06
    .MIN
    0.06
     legendary
    0.06
     './
    0.06
    proposal
    0.06
    Activations Density 0.032%

    No Known Activations