© 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 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. 45077
    Prev
    Next
    INDEX
    Explanations

    online forum posts

    np_max-act · gemini-2.0-flash

    The neuron is highly tuned to the protocol‐level formatting tokens that mark message headers (e.g. “<|start_header_id|>” and “<|end_header_id|>”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    conversational cues and social context in dialogue.

    oai_token-act-pair · gpt-4o-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
     elites
    -0.06
    (Direction
    -0.06
     制
    -0.06
    oran
    -0.06
     Spaces
    -0.06
    )'),↵
    -0.06
    (Print
    -0.06
    _import
    -0.06
     FILE
    -0.06
     др
    -0.06
    POSITIVE LOGITS
     заход
    0.09
    raz
    0.07
     mView
    0.07
    TableRow
    0.06
     život
    0.06
    _dd
    0.06
    .pictureBox
    0.06
     rencontrer
    0.06
     Tata
    0.06
    итор
    0.06
    Activations Density 0.069%

    No Known Activations