© 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. 8015
    Prev
    Next
    INDEX
    Explanations

    expressions of self-awareness and self-reflection

    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
    490
    +0.10
    0.5%
    602
    +0.10
    0.5%
    20
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    323
    +0.10
    0.04
    394
    +0.10
    0.03
    741
    +0.09
    0.03
    Negative Logits
     occupancy
    -0.71
    yip
    -0.66
     GOODMAN
    -0.66
    ondo
    -0.63
     migr
    -0.61
    hops
    -0.61
    ochet
    -0.61
    osate
    -0.60
    xit
    -0.60
    oping
    -0.60
    POSITIVE LOGITS
     brav
    0.75
    aunted
    0.72
    itled
    0.71
     pride
    0.70
     mischief
    0.67
     admiration
    0.67
     humility
    0.66
     modesty
    0.65
     vengeance
    0.65
     remorse
    0.64
    Activations Density 0.298%

    No Known Activations