© 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. 6105
    Prev
    Next
    INDEX
    Explanations

    proper nouns, specifically names of individuals in a musical context

    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
    +0.16
    0.8%
    106
    +0.13
    0.6%
    302
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.16
    0.06
    659
    +0.13
    0.06
    106
    +0.11
    0.04
    Negative Logits
     conv
    -0.66
    lli
    -0.63
    20439
    -0.63
    Ĥª
    -0.61
    ople
    -0.60
    potion
    -0.60
     manufact
    -0.60
     multipl
    -0.60
    naires
    -0.59
     erad
    -0.59
    POSITIVE LOGITS
     Jae
    1.16
    sung
    1.11
     Sung
    1.03
     Chung
    1.03
    jong
    1.03
    ongyang
    1.01
     Jong
    1.00
     Suk
    1.00
     Jung
    1.00
    jin
    1.00
    Activations Density 0.372%

    No Known Activations