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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 367
    Prev
    Next
    INDEX
    Explanations

    legal and technical terms related to a specific field, potentially related to cryptocurrency or technology

    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
    752
    +0.12
    0.5%
    68
    +0.12
    0.5%
    216
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.17
    455
    +0.12
    0.12
    40
    +0.11
    0.10
    Negative Logits
     Dame
    -0.74
    bones
    -0.73
    ÄŁ
    -0.68
    ILLE
    -0.65
     Cathedral
    -0.65
    bats
    -0.63
    fficiency
    -0.60
    ANCE
    -0.57
    edes
    -0.57
    ãĥīãĥ©
    -0.57
    POSITIVE LOGITS
    naires
    0.97
    naire
    0.91
    als
    0.82
    ages
    0.80
    atted
    0.76
    icals
    0.75
    ional
    0.73
     of
    0.72
    aler
    0.72
    english
    0.72
    Activations Density 2.813%

    No Known Activations