© 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. 6445
    Prev
    Next
    INDEX
    Explanations

    references to terminals and their associated characteristics or conditions

    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
    288
    +0.14
    0.6%
    515
    +0.12
    0.5%
    114
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.03
    515
    +0.12
    0.02
    211
    +0.11
    0.01
    Negative Logits
    akening
    -0.82
    hari
    -0.76
    ork
    -0.74
    ski
    -0.71
    ered
    -0.70
    artisan
    -0.69
    anson
    -0.67
    sky
    -0.66
    hs
    -0.66
    eree
    -0.65
    POSITIVE LOGITS
     emulator
    0.94
    idad
    0.69
    fo
    0.68
    iary
    0.67
    POS
    0.66
     Phone
    0.65
     velocity
    0.63
    atoes
    0.62
    itone
    0.61
    igmatic
    0.61
    Activations Density 0.138%

    No Known Activations