© 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. 10-RES_POST_32K-OAI
    11. 32213
    Prev
    Next
    INDEX
    Explanations

    references to political leaders and their relationships

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    764
    +0.10
    0.5%
    204
    +0.09
    0.4%
    373
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    700
    +0.10
    0.08
    502
    +0.09
    0.06
    224
    +0.09
    0.07
    Negative Logits
    poon
    -0.77
     destiny
    -0.63
     nowhere
    -0.62
     Ruin
    -0.61
     compulsion
    -0.60
    76561
    -0.59
     intolerable
    -0.59
     untold
    -0.57
    shapeshifter
    -0.57
     trivial
    -0.57
    POSITIVE LOGITS
     attends
    1.19
     reacts
    1.13
     greets
    1.11
     gestures
    1.11
     speaks
    1.07
     listens
    1.07
     poses
    1.07
     celebrates
    1.06
     participates
    1.06
    window
    1.02
    Activations Density 0.619%

    No Known Activations