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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 26618
    Prev
    Next
    INDEX
    Explanations

    instances of the word "book" and its variations, indicating references to reservations or scheduled events

    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
    664
    +0.10
    0.5%
    343
    +0.10
    0.4%
    113
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    222
    +0.10
    0.01
    705
    +0.10
    0.01
    370
    +0.09
    0.01
    Negative Logits
    ple
    -0.80
    psc
    -0.73
    Downloadha
    -0.73
    atism
    -0.72
     VIDEOS
    -0.72
    ï¸ı
    -0.71
    otive
    -0.68
    akia
    -0.67
    peria
    -0.66
    psy
    -0.65
    POSITIVE LOGITS
     booked
    0.88
     flights
    0.88
     booking
    0.81
    Kings
    0.81
     appointments
    0.79
     hotels
    0.73
     reservations
    0.69
     cancell
    0.69
     Reserv
    0.68
     tickets
    0.68
    Activations Density 0.011%

    No Known Activations