© 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. 6881
    Prev
    Next
    INDEX
    Explanations

    transitional phrases related to movement or change

    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
    165
    +0.10
    0.5%
    490
    +0.10
    0.4%
    257
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.10
    0.07
    123
    +0.10
    0.05
    362
    +0.09
    0.05
    Negative Logits
    ensis
    -0.77
     moder
    -0.73
    til
    -0.71
    ãĥ¼ãĤ¯
    -0.70
    haul
    -0.69
    cliffe
    -0.68
    ahime
    -0.67
     til
    -0.66
     calibrated
    -0.65
     conditioning
    -0.64
    POSITIVE LOGITS
     tears
    0.86
     laughter
    0.82
     gunfire
    0.77
    ocial
    0.69
    Bang
    0.66
     spontaneous
    0.66
    arious
    0.65
     surprise
    0.63
     joyful
    0.63
    ENCE
    0.62
    Activations Density 0.435%

    No Known Activations