© 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. 6428
    Prev
    Next
    INDEX
    Explanations

    references to tongues and ropes in various contexts

    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
    639
    +0.11
    0.5%
    23
    +0.10
    0.5%
    102
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.11
    0.03
    288
    +0.10
    0.03
    737
    +0.10
    0.02
    Negative Logits
    eanor
    -0.86
    rison
    -0.84
    ilion
    -0.84
    olitan
    -0.83
    icipated
    -0.79
    ysical
    -0.79
    ramid
    -0.75
    rons
    -0.73
    rano
    -0.72
    schild
    -0.71
    POSITIVE LOGITS
    Assembly
    0.80
    edge
    0.77
    gling
    0.69
     lash
    0.68
     hitch
    0.66
     Strikes
    0.65
     protr
    0.65
     craw
    0.64
    ength
    0.63
    á½
    0.63
    Activations Density 0.109%

    No Known Activations