© 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. 10-RES_POST_32K-OAI
    11. 32380
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Of" and its variations to indicate references or relationships in a context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    314
    +0.12
    0.5%
    333
    +0.12
    0.5%
    61
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    90
    +0.12
    0.03
    634
    +0.12
    0.03
    585
    +0.09
    0.03
    Negative Logits
     ashore
    -0.71
     evolves
    -0.69
     behaves
    -0.67
     dismant
    -0.61
     behave
    -0.60
     sails
    -0.60
     slump
    -0.59
     violates
    -0.59
    soType
    -0.58
     istg
    -0.57
    POSITIVE LOGITS
    Ļ½
    1.04
     note
    0.96
     particular
    0.94
    sted
    0.85
    cest
    0.84
    icial
    0.79
    µ
    0.79
    drawn
    0.78
     notable
    0.78
    ounding
    0.77
    Activations Density 0.142%

    No Known Activations