© 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. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 7347
    Prev
    Next
    INDEX
    Explanations

    mentions of servants and related roles in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    36
    +0.12
    0.5%
    288
    +0.11
    0.5%
    415
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.12
    0.06
    288
    +0.11
    0.06
    321
    +0.09
    0.04
    Negative Logits
    gments
    -0.86
    itar
    -0.82
    zee
    -0.79
    nesota
    -0.77
    inez
    -0.77
    emi
    -0.76
    orb
    -0.76
    ning
    -0.76
    Matrix
    -0.74
    edIn
    -0.74
    POSITIVE LOGITS
     servants
    0.88
     servant
    0.75
     maid
    0.74
     spoil
    0.73
     delinqu
    0.72
     suite
    0.72
     hall
    0.68
     Huss
    0.68
     Gardens
    0.67
     whisk
    0.66
    Activations Density 0.380%

    No Known Activations