© 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. 8596
    Prev
    Next
    INDEX
    Explanations

    references to the color orange or items associated with the color orange

    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
    78
    +0.11
    0.5%
    603
    +0.09
    0.4%
    445
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    78
    +0.11
    0.01
    603
    +0.09
    0.01
    433
    +0.09
    0.01
    Negative Logits
    enance
    -0.82
    rets
    -0.79
    ted
    -0.72
    Ö¼
    -0.72
    bler
    -0.70
    stood
    -0.69
    rior
    -0.69
    tle
    -0.67
    ts
    -0.67
    atted
    -0.66
    POSITIVE LOGITS
     peel
    1.08
     Blossom
    0.99
     juice
    0.96
    vale
    0.94
    issance
    0.92
     Peel
    0.89
     Juice
    0.82
     soup
    0.76
    ciation
    0.74
     cones
    0.70
    Activations Density 0.034%

    No Known Activations