© 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. 8561
    Prev
    Next
    INDEX
    Explanations

    instances of the word "lack" or its variations, indicating deficiencies or shortcomings

    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
    455
    +0.12
    0.5%
    566
    +0.11
    0.5%
    633
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    455
    +0.12
    0.03
    641
    +0.11
    0.02
    234
    +0.11
    0.02
    Negative Logits
    è¦ļéĨĴ
    -0.80
    shirt
    -0.76
    MG
    -0.75
    rolled
    -0.75
    ping
    -0.72
    mad
    -0.71
    terms
    -0.71
    uum
    -0.71
    uo
    -0.71
    kind
    -0.70
    POSITIVE LOGITS
     empathy
    0.99
     specificity
    0.95
     patience
    0.94
     nuance
    0.91
     resources
    0.90
     credibility
    0.90
     cohesion
    0.88
     adequate
    0.87
     clout
    0.87
     stamina
    0.87
    Activations Density 0.100%

    No Known Activations