© 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 EvalsExportsAPI 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. 6923
    Prev
    Next
    INDEX
    Explanations

    specific symbols, characters, or formatting typically associated with data representations and citations

    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
    524
    +0.12
    0.5%
    365
    +0.10
    0.5%
    160
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    675
    +0.12
    0.03
    740
    +0.10
    0.02
    524
    +0.10
    0.03
    Negative Logits
    paces
    -0.79
    yip
    -0.72
    ï¸ı
    -0.68
    ngth
    -0.68
    soever
    -0.67
     cruise
    -0.67
     bombers
    -0.66
    gradation
    -0.66
     lame
    -0.66
     framed
    -0.63
    POSITIVE LOGITS
    Consumer
    0.86
     Cosponsors
    0.86
    English
    0.84
    Capital
    0.81
    United
    0.77
    Legendary
    0.75
    Technology
    0.73
    Education
    0.73
    rg
    0.71
    Unity
    0.70
    Activations Density 0.120%

    No Known Activations