© 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. 10-RES_POST_32K-OAI
    11. 30099
    Prev
    Next
    INDEX
    Explanations

    references to detectives or detective-related activities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    196
    +0.12
    0.6%
    477
    +0.12
    0.5%
    283
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.12
    0.01
    196
    +0.12
    0.01
    178
    +0.10
    0.01
    Negative Logits
    manship
    -0.71
     Nicarag
    -0.65
     freeze
    -0.63
     Tobacco
    -0.61
     gesture
    -0.60
    #$
    -0.60
    lihood
    -0.60
     bold
    -0.60
     centimeters
    -0.59
     deduct
    -0.58
    POSITIVE LOGITS
    achable
    1.41
    achment
    1.28
    ailed
    1.28
    roit
    1.24
    ector
    1.23
    ective
    1.21
    ached
    1.19
    our
    1.13
    rit
    1.13
    ainer
    1.13
    Activations Density 0.055%

    No Known Activations