© 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. 6304
    Prev
    Next
    INDEX
    Explanations

    elements related to statistical analysis and control variables in research studies

    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
    408
    +0.09
    0.4%
    763
    +0.09
    0.4%
    511
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    408
    +0.09
    0.15
    236
    +0.09
    0.11
    724
    +0.08
    0.12
    Negative Logits
    Led
    -0.83
    isode
    -0.76
    å§
    -0.74
    heading
    -0.72
    scribe
    -0.72
    ileaks
    -0.72
    tumblr
    -0.71
     largeDownload
    -0.70
    ãĥĺ
    -0.70
    named
    -0.69
    POSITIVE LOGITS
     uncertainties
    1.25
     factors
    1.22
     biases
    1.18
     confounding
    1.16
     variables
    1.02
     covari
    1.01
     considerations
    1.01
     constraints
    1.00
     heterogeneity
    0.99
     methodological
    0.99
    Activations Density 3.065%

    No Known Activations