© 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. 6987
    Prev
    Next
    INDEX
    Explanations

    verbs and phrases that indicate reference or relation to concepts

    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
    +0.12
    0.5%
    575
    +0.11
    0.5%
    164
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    195
    +0.12
    0.02
    448
    +0.11
    0.02
    759
    +0.11
    0.02
    Negative Logits
    dar
    -0.68
     Carbuncle
    -0.60
     helicop
    -0.59
     Moroccan
    -0.58
     Muk
    -0.57
    haw
    -0.57
     corpus
    -0.57
     Louie
    -0.56
     Croat
    -0.55
     Nar
    -0.54
    POSITIVE LOGITS
     principally
    0.96
     primarily
    0.91
     chiefly
    0.90
    entially
    0.90
     uniquely
    0.88
     exclusively
    0.87
     solely
    0.84
     directly
    0.84
     specifically
    0.82
    rough
    0.81
    Activations Density 0.089%

    No Known Activations