Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 177
    Prev
    Next
    INDEX
    Explanations

    phrases related to secrecy or hiding

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    288
    +0.13
    0.6%
    722
    +0.10
    0.5%
    67
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.14
    1
    +0.10
    0.07
    592
    +0.10
    0.06
    Negative Logits
    ctive
    -0.71
    oday
    -0.71
    baugh
    -0.71
    reciation
    -0.70
    nesota
    -0.69
    rompt
    -0.69
    ucc
    -0.69
    aldi
    -0.69
    annis
    -0.69
    ragon
    -0.67
    POSITIVE LOGITS
     secrets
    0.98
     confidential
    0.90
     hidden
    0.89
     Secrets
    0.88
    ariat
    0.87
    arial
    0.87
     cloaked
    0.85
     informant
    0.85
     secret
    0.84
     stash
    0.81
    Activations Density 2.543%

    No Known Activations