© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 19949
    Prev
    Next
    INDEX
    Explanations

    movies and TV show references, especially related to costumes and storylines

    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
    567
    +0.12
    0.5%
    525
    +0.10
    0.5%
    248
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.12
    0.04
    652
    +0.10
    0.03
    567
    +0.10
    0.02
    Negative Logits
    xual
    -0.96
    ilial
    -0.77
    ulty
    -0.76
    aucuses
    -0.74
    amination
    -0.73
    indu
    -0.72
    yll
    -0.70
    hematically
    -0.70
    ĸļ
    -0.69
    oxicity
    -0.69
    POSITIVE LOGITS
     groceries
    1.13
     tickets
    1.01
     cheapest
    0.96
     discounted
    0.94
     Tickets
    0.89
     cheaply
    0.89
     cheap
    0.88
     licences
    0.87
     pricey
    0.86
     goods
    0.85
    Activations Density 0.183%

    No Known Activations