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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 126309
    Prev
    Next
    INDEX
    Explanations

    -d

    np_max-act · gemini-2.0-flash

    The neuron specifically fires on the “-d” fragment in the “class-dump” header lines, i.e. it detects the “-d” token appearing in the autogenerated class-dump comments.

    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
    ____________
    -0.07
    (IO
    -0.07
     seal
    -0.07
     kuvvet
    -0.07
    алом
    -0.07
    endi
    -0.07
    ToLocal
    -0.06
    (encoding
    -0.06
    /img
    -0.06
     Kuwait
    -0.06
    POSITIVE LOGITS
     regards
    0.06
     دیگر
    0.06
     situation
    0.06
     drum
    0.06
    )b
    0.06
    	cerr
    0.06
     driv
    0.06
    із
    0.06
     nichts
    0.06
    ειο
    0.06
    Activations Density 0.000%

    No Known Activations