© 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. 7892
    Prev
    Next
    INDEX
    Explanations

    instances of the word "by" in various contexts, indicating a focus on agency and attribution

    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
    669
    +0.17
    0.8%
    608
    +0.14
    0.6%
    362
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.17
    0.08
    669
    +0.14
    0.07
    362
    +0.13
    0.06
    Negative Logits
    osen
    -0.77
    emetery
    -0.77
    iatrics
    -0.76
    secut
    -0.75
    eas
    -0.73
    wrong
    -0.73
    fox
    -0.72
    gue
    -0.70
    iate
    -0.70
    mint
    -0.69
    POSITIVE LOGITS
     visions
    1.02
     fears
    0.92
     warnings
    0.86
     fear
    0.80
     tales
    0.79
     rumours
    0.79
     messages
    0.78
     stories
    0.78
     advances
    0.77
     doubts
    0.76
    Activations Density 0.351%

    No Known Activations