© 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. 7669
    Prev
    Next
    INDEX
    Explanations

    key terms related to identification and assessment of individuals or entities

    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
    497
    +0.11
    0.5%
    692
    +0.10
    0.5%
    429
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.11
    0.08
    393
    +0.10
    0.07
    608
    +0.10
    0.07
    Negative Logits
     reckoning
    -0.80
    ocaust
    -0.73
    advertising
    -0.69
    finder
    -0.67
     bashing
    -0.64
    env
    -0.64
    ranging
    -0.63
    depth
    -0.63
    estate
    -0.62
    fortune
    -0.62
    POSITIVE LOGITS
     culprit
    1.25
     weaknesses
    0.93
     flaws
    0.89
     landmarks
    0.86
     vulnerabilities
    0.83
     deficiencies
    0.81
     flaw
    0.80
     fault
    0.79
     perpetrators
    0.78
     perpetrator
    0.77
    Activations Density 0.634%

    No Known Activations