© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 15-RESID-POST-AA
    11. 87294
    Prev
    Next
    INDEX
    Explanations

    th

    np_max-act · gemini-2.0-flash

    tokens related to document structure or formatting and prompt-style instructions (headings, hashes, metadata/instruction cues).

    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
    	tmp
    -0.07
     Combo
    -0.07
    　l
    -0.06
    _series
    -0.06
    [::-
    -0.06
    _surf
    -0.06
     нагруз
    -0.06
     forgive
    -0.06
     tear
    -0.06
    .Span
    -0.06
    POSITIVE LOGITS
    ```
    0.08
     приготовить
    0.07
    /QĐ
    0.06
    ##
    0.06
     Coconut
    0.06
     Зав
    0.06
    ристи
    0.06
    くん
    0.06
    ;↵↵↵
    0.06
     Hern
    0.06
    Activations Density 0.000%

    No Known Activations