© 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. 8-RES_POST_32K-OAI
    11. 6761
    Prev
    Next
    INDEX
    Explanations

    sequences of underscores and colons commonly seen in code or configuration contexts

    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
    479
    +0.10
    0.5%
    741
    +0.10
    0.5%
    91
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.10
    0.03
    107
    +0.10
    0.03
    741
    +0.10
    0.02
    Negative Logits
    multipl
    -1.14
    matter
    -1.12
    pron
    -1.03
    techn
    -0.99
    need
    -0.99
    practice
    -0.97
    ever
    -0.96
    tein
    -0.96
    indal
    -0.95
    cause
    -0.94
    POSITIVE LOGITS
     XI
    0.81
     Huck
    0.79
     XV
    0.77
     Bust
    0.69
     Ashes
    0.66
     XVI
    0.66
     Majesty
    0.65
     Kush
    0.65
     summers
    0.65
     Lowry
    0.65
    Activations Density 0.090%

    No Known Activations