© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 4956
    Prev
    Next
    INDEX
    Explanations

    numerical values, especially related to addresses or locations

    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
    +0.13
    0.6%
    538
    +0.09
    0.4%
    217
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.13
    0.03
    689
    +0.09
    0.03
    260
    +0.09
    0.03
    Negative Logits
    icter
    -0.78
    accompan
    -0.74
    lied
    -0.69
    otrop
    -0.68
    doms
    -0.67
     seiz
    -0.66
     flavours
    -0.66
    successfully
    -0.66
    aper
    -0.65
     destro
    -0.65
    POSITIVE LOGITS
     NW
    1.06
     SW
    0.97
     Greenwich
    0.95
     Broadway
    0.94
     East
    0.94
     Rue
    0.93
     Antioch
    0.92
     Arlington
    0.92
     Sunset
    0.91
     Riverside
    0.90
    Activations Density 0.141%

    No Known Activations