© 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. 5029
    Prev
    Next
    INDEX
    Explanations

    symbols and non-standard characters

    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
    373
    +0.22
    1.0%
    244
    +0.11
    0.5%
    207
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    107
    +0.22
    0.17
    740
    +0.11
    0.15
    514
    +0.09
    0.14
    Negative Logits
    packing
    -0.69
    */(
    -0.62
     cle
    -0.60
    wcs
    -0.59
    phia
    -0.59
    paces
    -0.58
    issance
    -0.57
     horizont
    -0.56
     carrier
    -0.55
     contingency
    -0.55
    POSITIVE LOGITS
    ----------------------------------------------------------------
    0.76
    --------------------------------
    0.74
    --------------------------------------------------------
    0.74
    ________________
    0.74
    ********************************
    0.66
    Available
    0.66
    acters
    0.65
    ------
    0.64
    ================================
    0.64
    efully
    0.64
    Activations Density 1.412%

    No Known Activations