© 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. 29133
    Prev
    Next
    INDEX
    Explanations

    conjunctions and their variations highlighting connections between ideas

    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
    30
    +0.10
    0.5%
    512
    +0.08
    0.4%
    290
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    239
    +0.10
    0.02
    290
    +0.08
    0.02
    631
    +0.08
    0.01
    Negative Logits
     Journals
    -0.80
    uously
    -0.65
     Kard
    -0.65
    Byte
    -0.65
     Chains
    -0.65
     Breed
    -0.61
    IUM
    -0.61
    leaf
    -0.60
    Fle
    -0.60
    EH
    -0.60
    POSITIVE LOGITS
    during
    0.74
     during
    0.71
     aft
    0.68
     contempor
    0.64
    effects
    0.64
    ologue
    0.62
    rogen
    0.61
    eret
    0.59
    aeda
    0.58
    behind
    0.58
    Activations Density 0.205%

    No Known Activations