© 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. 68430
    Prev
    Next
    INDEX
    Explanations

    code snippets

    np_max-act · gemini-2.0-flash

    The neuron activates on numeric version‐style tokens and other standalone numbers (e.g. “1.0,” “2007,” “254,” etc.), effectively spotting version numbers and numeric literals.

    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
    EQUAL
    -0.07
     Eduardo
    -0.07
    bfd
    -0.06
     competing
    -0.06
    GLE
    -0.06
    wend
    -0.05
    OFF
    -0.05
    REATE
    -0.05
    -design
    -0.05
     detrimental
    -0.05
    POSITIVE LOGITS
    (Control
    0.07
     الاع
    0.07
     ar
    0.07
     gitti
    0.07
    ******↵↵
    0.07
     ар
    0.07
     đáo
    0.07
    [jj
    0.07
    	db
    0.07
    )、
    0.07
    Activations Density 0.004%

    No Known Activations