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

    Neuronpedia

    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. 4297
    Prev
    Next
    INDEX
    Explanations

    references to the word "se."

    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
    631
    +0.11
    0.5%
    110
    +0.10
    0.5%
    696
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    631
    +0.11
    0.03
    696
    +0.10
    0.03
    599
    +0.09
    0.03
    Negative Logits
     banter
    -0.67
     Thumbnails
    -0.66
     AFL
    -0.66
     diplom
    -0.64
     BDS
    -0.62
     nickel
    -0.61
    cake
    -0.60
    stick
    -0.59
     Judd
    -0.59
     overhead
    -0.58
    POSITIVE LOGITS
    eking
    1.23
    venth
    1.22
    ldom
    1.19
    lected
    1.18
    ems
    1.17
    quel
    1.15
    gments
    1.13
    vere
    1.12
    vered
    1.10
    eker
    1.08
    Activations Density 0.077%

    No Known Activations