© 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. 8401
    Prev
    Next
    INDEX
    Explanations

    phrases associated with searching or seeking solutions

    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
    665
    +0.10
    0.5%
    690
    +0.09
    0.4%
    133
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.10
    0.04
    665
    +0.09
    0.06
    604
    +0.09
    0.05
    Negative Logits
    minus
    -0.67
    forge
    -0.64
     disbelief
    -0.63
     reeling
    -0.63
    gd
    -0.62
    panic
    -0.61
    eka
    -0.61
    ismo
    -0.60
     Appears
    -0.59
    rylic
    -0.58
    POSITIVE LOGITS
     ways
    1.14
     effic
    0.93
     solutions
    0.87
     optimal
    0.85
     loopholes
    0.80
     alternatives
    0.80
     replacements
    0.77
     criteria
    0.77
     synerg
    0.76
     suitable
    0.76
    Activations Density 0.334%

    No Known Activations