© 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. 11-RESID-POST-AA
    11. 58289
    Prev
    Next
    INDEX
    Explanations

    conflicts or duplicates

    np_max-act · gemini-2.0-flash

    This neuron detects snippets of technical error or exception messages (e.g. “already the logical child of another element,” stack‐trace lines, repeated property‐set warnings).

    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
     sons
    -0.07
     genome
    -0.07
    -mask
    -0.06
    _inner
    -0.06
     tote
    -0.06
     Gentle
    -0.06
     Simply
    -0.06
    assword
    -0.06
     astronom
    -0.06
     loop
    -0.06
    POSITIVE LOGITS
    foregroundColor
    0.07
    ]‏
    0.07
    __((
    0.07
    Animating
    0.07
    NPC
    0.06
    (MainActivity
    0.06
    _KeyPress
    0.06
     adulte
    0.06
     NPCs
    0.06
     клу
    0.06
    Activations Density 0.039%

    No Known Activations