© 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. 4240
    Prev
    Next
    INDEX
    Explanations

    references to political figures and their titles

    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
    150
    +0.10
    0.5%
    728
    +0.10
    0.4%
    526
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.10
    0.03
    224
    +0.10
    0.03
    418
    +0.09
    0.03
    Negative Logits
    ally
    -0.99
    ioned
    -0.84
    ansas
    -0.81
    rarily
    -0.81
    aska
    -0.80
     Tribe
    -0.78
    apon
    -0.78
    LOAD
    -0.74
    ions
    -0.74
    ibaba
    -0.72
    POSITIVE LOGITS
    llor
    1.25
     Merkel
    1.10
     Chancellor
    0.88
     chancellor
    0.73
    swer
    0.73
    ãĥ´ãĤ¡
    0.69
     Nero
    0.67
     Lans
    0.65
    ylum
    0.64
    itzer
    0.64
    Activations Density 0.133%

    No Known Activations