© 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. 10-RES_POST_32K-OAI
    11. 28145
    Prev
    Next
    INDEX
    Explanations

    numeric values associated with statistical data

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    303
    +0.10
    0.5%
    314
    +0.10
    0.5%
    315
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.10
    0.04
    316
    +0.10
    0.04
    96
    +0.10
    0.03
    Negative Logits
    alia
    -0.71
    iple
    -0.64
     inval
    -0.63
    aneously
    -0.58
     objectionable
    -0.57
     imped
    -0.57
     drawbacks
    -0.57
     away
    -0.56
    enn
    -0.55
     overboard
    -0.55
    POSITIVE LOGITS
    001
    1.39
    0001
    1.21
    05
    1.21
    01
    1.09
    005
    1.00
    0000
    0.97
    0002
    0.96
    002
    0.90
    018
    0.90
    011
    0.88
    Activations Density 0.223%

    No Known Activations