© 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. 32236
    Prev
    Next
    INDEX
    Explanations

    words related to repetition or reiteration

    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
    280
    +0.10
    0.4%
    357
    +0.10
    0.4%
    260
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    701
    +0.10
    0.02
    314
    +0.10
    -0.00
    232
    +0.09
    0.03
    Negative Logits
     Unch
    -0.61
     Cornel
    -0.59
     arms
    -0.58
     plat
    -0.57
     bott
    -0.56
     z
    -0.56
     Panda
    -0.56
     shoulders
    -0.56
    hurst
    -0.55
     Polly
    -0.54
    POSITIVE LOGITS
    igent
    0.93
    lycer
    0.88
    uders
    0.84
    idate
    0.84
    usive
    0.78
    spection
    0.77
    uration
    0.74
    ngth
    0.74
    acies
    0.74
    LECT
    0.74
    Activations Density 0.181%

    No Known Activations