© 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 EvalsExportsinterp-engineNEWAPI 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. 8345
    Prev
    Next
    INDEX
    Explanations

    references to specific groups or categories

    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
    540
    +0.10
    0.5%
    557
    +0.10
    0.5%
    497
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1
    +0.10
    0.04
    45
    +0.10
    0.04
    450
    +0.09
    0.04
    Negative Logits
    endment
    -0.71
     hypocrisy
    -0.70
     novelist
    -0.66
    ocratic
    -0.65
    vier
    -0.63
     nationalist
    -0.62
     protagonist
    -0.61
    eur
    -0.61
    cellence
    -0.61
     psychiatrist
    -0.61
    POSITIVE LOGITS
     contain
    0.91
     are
    0.81
    atically
    0.80
     originated
    0.80
     differs
    0.80
     vary
    0.80
     belong
    0.78
     differ
    0.78
     relate
    0.78
     may
    0.76
    Activations Density 0.174%

    No Known Activations