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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 41354
    Prev
    Next
    INDEX
    Explanations

    references to the term "Blues" associated with a sports team or context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    8
    +0.11
    0.5%
    290
    +0.10
    0.5%
    649
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    269
    +0.11
    0.01
    173
    +0.10
    0.01
    290
    +0.10
    0.01
    Negative Logits
    awei
    -0.81
    igators
    -0.72
    rises
    -0.71
    igated
    -0.69
    ADRA
    -0.67
    ortion
    -0.67
    ILY
    -0.67
    ilings
    -0.65
     Democr
    -0.65
     Gingrich
    -0.64
    POSITIVE LOGITS
    hift
    1.36
    hirt
    1.16
    creen
    1.05
    pace
    0.84
    ignt
    0.83
    keeper
    0.79
    berries
    0.75
    wagen
    0.75
    keepers
    0.74
     Blues
    0.71
    Activations Density 0.005%

    No Known Activations