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. 278
    Prev
    Next
    INDEX
    Explanations

    terms related to quantity

    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
    71
    +0.16
    0.7%
    662
    +0.11
    0.5%
    8
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    71
    +0.16
    0.13
    344
    +0.11
    0.14
    726
    +0.10
    0.12
    Negative Logits
    opia
    -0.73
    atum
    -0.72
    acus
    -0.70
    istan
    -0.70
    heed
    -0.68
    ivism
    -0.67
    UAL
    -0.66
    agame
    -0.66
    ensis
    -0.66
    icism
    -0.66
    POSITIVE LOGITS
     facets
    1.13
     times
    1.13
     occasions
    1.09
     aspects
    1.03
     instances
    1.02
     thousands
    0.98
     people
    0.94
     unanswered
    0.93
     others
    0.92
     jurisdictions
    0.92
    Activations Density 2.609%

    No Known Activations