© 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 49152
    9. 8-RES_FS49152-JB
    11. 41669
    Prev
    Next
    INDEX
    Explanations

    terms related to voice acting and audio production

    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
    447
    +0.13
    0.6%
    83
    +0.11
    0.5%
    679
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    401
    +0.13
    0.02
    505
    +0.11
    0.02
    83
    +0.09
    0.01
    Negative Logits
    cot
    -0.85
     Klux
    -0.77
    erenn
    -0.75
    ldon
    -0.75
    bard
    -0.73
    ndra
    -0.71
     Citiz
    -0.70
    Vert
    -0.70
    ecause
    -0.70
    ulner
    -0.69
    POSITIVE LOGITS
     narration
    1.29
     subtitles
    1.21
     announcer
    1.09
     vocals
    1.03
     interpreter
    1.00
     voice
    0.99
    Voice
    0.96
     streamed
    0.92
     recordings
    0.92
     louder
    0.92
    Activations Density 0.065%

    No Known Activations