© 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. 6384
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing the concept of "more."

    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
    557
    +0.13
    0.6%
    500
    +0.10
    0.5%
    491
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    389
    +0.13
    0.03
    726
    +0.10
    0.03
    13
    +0.10
    0.03
    Negative Logits
    ities
    -0.83
    ity
    -0.77
    edom
    -0.72
    sheet
    -0.72
    thood
    -0.70
    acters
    -0.67
    oké
    -0.66
    ighting
    -0.66
    monary
    -0.66
    lihood
    -0.65
    POSITIVE LOGITS
     abundantly
    0.68
     thankful
    0.67
     frustrating
    0.67
    Unt
    0.66
    SOURCE
    0.64
     grateful
    0.61
     succinct
    0.61
    GBT
    0.60
     thrilled
    0.60
     stoked
    0.60
    Activations Density 0.124%

    No Known Activations