© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4553
    Prev
    Next
    INDEX
    Explanations

    notations or markers used in coding or programming documents

    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
    640
    +0.10
    0.5%
    393
    +0.10
    0.5%
    307
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    176
    +0.10
    0.02
    6
    +0.10
    0.03
    742
    +0.09
    0.02
    Negative Logits
    æ©
    -0.77
    boro
    -0.73
    bike
    -0.67
    bledon
    -0.67
    omething
    -0.65
    bent
    -0.61
    houses
    -0.60
    APS
    -0.60
    Tube
    -0.60
    issan
    -0.59
    POSITIVE LOGITS
     Detected
    1.20
     Adding
    0.99
     WARNING
    0.97
     Loading
    0.95
     =================================
    0.93
     Sending
    0.92
     Unable
    0.90
     Reload
    0.90
     Invalid
    0.89
     Checking
    0.88
    Activations Density 0.092%

    No Known Activations