© 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. 128527
    Prev
    Next
    INDEX
    Explanations

    formal documents

    np_max-act · gemini-2.0-flash

    The neuron activates on single‐letter capital tokens (often followed by punctuation) that signal author initials or similar metadata fields in academic paper headers.

    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
     miscellaneous
    -0.07
    .Interfaces
    -0.06
    _NAMESPACE
    -0.06
     تجربه
    -0.06
     Microsoft
    -0.06
     responsibility
    -0.06
    .getContent
    -0.06
     Couple
    -0.06
    wf
    -0.06
     Detector
    -0.06
    POSITIVE LOGITS
    	element
    0.06
     wich
    0.06
    dest
    0.06
    iv
    0.06
    (cal
    0.06
     Certain
    0.06
     Eagles
    0.06
    _large
    0.05
    roots
    0.05
     PLACE
    0.05
    Activations Density 0.150%

    No Known Activations