© 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. 8149
    Prev
    Next
    INDEX
    Explanations

    geographical locations and landmarks

    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
    586
    +0.10
    0.4%
    763
    +0.10
    0.4%
    721
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.10
    0.15
    288
    +0.10
    0.20
    585
    +0.09
    0.15
    Negative Logits
     disruptive
    -0.84
     upfront
    -0.77
     ObamaCare
    -0.76
     boycot
    -0.75
     hurting
    -0.72
     arbitration
    -0.72
    blocking
    -0.71
    amacare
    -0.71
    calling
    -0.70
     pressuring
    -0.69
    POSITIVE LOGITS
     Museum
    1.24
     archae
    1.23
     archaeologists
    1.17
     curator
    1.16
     Archae
    1.15
     Hist
    1.08
     historian
    1.06
     museum
    1.05
    useum
    1.03
     Historical
    1.02
    Activations Density 5.453%

    No Known Activations