© 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. 12476
    Prev
    Next
    INDEX
    Explanations

    contractions

    np_max-act · gemini-2.0-flash

    This neuron detects speaker‐role or header tokens (e.g. the “assistant” tag) in the dialogue format.

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

    the assistant role header marking the start of an assistant message in chat-style transcripts.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
    -console
    -0.07
    >#
    -0.07
    .ping
    -0.07
     uncommon
    -0.06
    てい
    -0.06
     görül
    -0.06
    Viewer
    -0.06
    ampilkan
    -0.06
    ih
    -0.06
     xhr
    -0.06
    POSITIVE LOGITS
    ilarity
    0.06
    _dw
    0.06
     classification
    0.06
     necessity
    0.06
     fired
    0.06
     vant
    0.06
     susp
    0.06
    Scan
    0.06
    Jet
    0.06
    ity
    0.06
    Activations Density 0.092%

    No Known Activations