© 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. 8166
    Prev
    Next
    INDEX
    Explanations

    sections that detail authorship, contributions, and acknowledgments in academic papers

    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
    694
    +0.12
    0.5%
    82
    +0.10
    0.5%
    2
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    586
    +0.12
    0.04
    214
    +0.10
    0.03
    211
    +0.10
    0.02
    Negative Logits
     ItemLevel
    -0.75
     guiName
    -0.71
    ioxide
    -0.68
    gdala
    -0.67
    ouston
    -0.67
     [/
    -0.66
    agame
    -0.64
     Enlight
    -0.64
     Tsukuyomi
    -0.63
    ounter
    -0.63
    POSITIVE LOGITS
     manuscript
    0.84
     donors
    0.83
     Ethics
    0.79
     donor
    0.78
     faculty
    0.77
    ipal
    0.75
     guarant
    0.73
    doctoral
    0.72
     editorial
    0.71
     acknowled
    0.69
    Activations Density 0.627%

    No Known Activations