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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 360
    Prev
    Next
    INDEX
    Explanations

    information related to conferences, exhibitions, and workshops

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    677
    +0.11
    0.5%
    430
    +0.10
    0.4%
    310
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    677
    +0.11
    0.12
    115
    +0.10
    0.14
    310
    +0.09
    0.13
    Negative Logits
     instincts
    -0.79
     instinct
    -0.77
     mistakes
    -0.72
     cracks
    -0.70
     screws
    -0.70
     messing
    -0.69
     porous
    -0.69
     brute
    -0.69
     workaround
    -0.68
     looph
    -0.67
    POSITIVE LOGITS
     attendees
    1.50
     festivities
    1.33
     hosted
    1.17
     Attend
    1.16
     Participants
    1.15
     commemor
    1.14
     festival
    1.14
     Celebration
    1.12
     venue
    1.12
     attendance
    1.11
    Activations Density 7.521%

    No Known Activations