© 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. 7577
    Prev
    Next
    INDEX
    Explanations

    references to legal citations and procedural terms

    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%
    547
    +0.12
    0.5%
    420
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    426
    +0.13
    0.11
    352
    +0.12
    0.12
    23
    +0.09
    0.12
    Negative Logits
    staking
    -0.82
     Frie
    -0.75
    ezvous
    -0.74
    uca
    -0.73
    ounces
    -0.72
    schild
    -0.71
    inia
    -0.70
    itimate
    -0.70
    ãĥ¯ãĥ³
    -0.68
    naires
    -0.67
    POSITIVE LOGITS
     repealed
    0.89
     Squirrel
    0.81
     repeal
    0.74
     chapter
    0.73
     NRS
    0.73
     Legislature
    0.71
     typo
    0.71
     ILCS
    0.70
     amended
    0.70
     Revision
    0.69
    Activations Density 0.676%

    No Known Activations