© 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. 10-RES_POST_32K-OAI
    11. 30424
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and specific structural elements in the text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    139
    +0.10
    0.4%
    224
    +0.09
    0.4%
    202
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    134
    +0.10
    0.02
    198
    +0.09
    0.01
    290
    +0.09
    0.01
    Negative Logits
    ãĥĥãĥī
    -0.85
     Gleaming
    -0.84
    ãĤ¦ãĤ¹
    -0.77
    ĸļ
    -0.74
     srfAttach
    -0.67
     Puzzles
    -0.64
     Werewolf
    -0.62
    DragonMagazine
    -0.62
    imov
    -0.62
    arthed
    -0.62
    POSITIVE LOGITS
    ository
    0.70
    riber
    0.69
    abouts
    0.66
    iona
    0.66
    opic
    0.61
    ships
    0.61
    idav
    0.60
    lo
    0.59
     diplom
    0.59
    oner
    0.58
    Activations Density 0.113%

    No Known Activations