© 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. 7890
    Prev
    Next
    INDEX
    Explanations

    the action of "doing"

    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
    179
    +0.12
    0.5%
    573
    +0.11
    0.5%
    218
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    78
    +0.12
    0.04
    113
    +0.11
    0.04
    98
    +0.09
    0.04
    Negative Logits
    naire
    -0.76
    scribe
    -0.75
    upon
    -0.73
    eric
    -0.73
    urally
    -0.72
    ethy
    -0.72
    ulate
    -0.70
    exist
    -0.68
    oat
    -0.67
    ãĤ¨ãĥ«
    -0.67
    POSITIVE LOGITS
     fine
    0.77
     soaking
    0.74
     pul
    0.73
     alright
    0.71
     hypocr
    0.67
     hiding
    0.65
     lax
    0.63
     apologizing
    0.63
     okay
    0.63
     stellar
    0.62
    Activations Density 0.269%

    No Known Activations