© 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 EvalsExportsAPI 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. 6911
    Prev
    Next
    INDEX
    Explanations

    forms of the verb "to live" and related concepts about existence and awareness

    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
    745
    +0.12
    0.5%
    52
    +0.09
    0.4%
    394
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    406
    +0.12
    0.07
    314
    +0.09
    0.00
    514
    +0.09
    0.06
    Negative Logits
    entry
    -0.62
     disguise
    -0.60
     req
    -0.60
     bookmark
    -0.58
     Antar
    -0.57
     inhibitor
    -0.57
    colour
    -0.56
     displacement
    -0.56
     puppet
    -0.56
     relegation
    -0.56
    POSITIVE LOGITS
    IGN
    0.91
    iously
    0.81
    otyp
    0.80
    ignty
    0.80
    chers
    0.79
    esome
    0.77
    cher
    0.77
    zing
    0.77
    erred
    0.75
    ches
    0.75
    Activations Density 0.695%

    No Known Activations