© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 4981
    Prev
    Next
    INDEX
    Explanations

    references to footnotes and citations in text

    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
    739
    +0.14
    0.7%
    641
    +0.12
    0.5%
    373
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    719
    +0.14
    0.05
    107
    +0.12
    0.05
    514
    +0.10
    0.04
    Negative Logits
    thood
    -0.85
    creen
    -0.85
    onz
    -0.81
    swick
    -0.80
    manship
    -0.77
    ettings
    -0.74
    ofi
    -0.73
    ows
    -0.73
    omething
    -0.72
    otin
    -0.70
    POSITIVE LOGITS
    Footnote
    1.01
     Foot
    0.76
     footnote
    0.73
     Brief
    0.69
     accompanying
    0.68
     Figures
    0.67
    edly
    0.63
     omitted
    0.63
     Reason
    0.60
     Dungeon
    0.59
    Activations Density 0.223%

    No Known Activations