© 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. 6433
    Prev
    Next
    INDEX
    Explanations

    instructions related to data entry or input tasks

    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
    279
    +0.14
    0.6%
    411
    +0.11
    0.5%
    206
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.14
    0.02
    604
    +0.11
    0.02
    59
    +0.10
    0.02
    Negative Logits
    arers
    -0.74
     Hutch
    -0.73
    iferation
    -0.65
    âĶ
    -0.65
    auld
    -0.64
     Pigs
    -0.63
    gas
    -0.63
    rador
    -0.62
    Reilly
    -0.62
     exhibiting
    -0.61
    POSITIVE LOGITS
     passwords
    1.14
    codes
    0.96
     coordinates
    0.94
     names
    0.92
     details
    0.92
     password
    0.90
     username
    0.89
     PIN
    0.88
     credentials
    0.86
     commands
    0.84
    Activations Density 0.245%

    No Known Activations