© 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. 7599
    Prev
    Next
    INDEX
    Explanations

    periods, particularly at the end of sentences

    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
    373
    +0.20
    0.9%
    640
    +0.13
    0.6%
    28
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.20
    0.04
    409
    +0.13
    0.04
    513
    +0.11
    0.03
    Negative Logits
     erosion
    -0.70
     ordinary
    -0.70
    ħĭ
    -0.68
     refreshed
    -0.68
     meaningful
    -0.67
    TextColor
    -0.63
    ©¶æ
    -0.63
    «ĺ
    -0.63
     sooner
    -0.63
     reasonable
    -0.63
    POSITIVE LOGITS
    edu
    1.17
    net
    0.96
    uni
    0.96
    github
    0.89
    amazon
    0.89
    gio
    0.88
    org
    0.87
    mb
    0.85
    nl
    0.85
    demon
    0.82
    Activations Density 0.111%

    No Known Activations