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 24576
    5. 8-RES_FS24576-JB
    6. 803
    Prev
    Next
    INDEX
    Explanations

    references to a specific company or brand name, particularly 'Dreamcast'

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    575
    +0.12
    0.6%
    218
    +0.11
    0.5%
    414
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.12
    0.01
    343
    +0.11
    0.01
    52
    +0.09
    0.01
    Negative Logits
     Adin
    -0.77
     menstru
    -0.71
    assed
    -0.66
    anthrop
    -0.65
    cules
    -0.65
    ahn
    -0.62
     Mellon
    -0.62
    idges
    -0.61
    ahon
    -0.61
    ength
    -0.60
    POSITIVE LOGITS
    Hack
    1.21
    hack
    1.14
    tek
    1.02
     Chaser
    1.02
    liner
    0.99
    scape
    0.97
    Works
    0.96
    pad
    0.96
    buster
    0.90
    liners
    0.87
    Activations Density 0.015%

    No Known Activations