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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 23935
    Prev
    Next
    INDEX
    Explanations

    narrative elements involving character development and plot twists

    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
    87
    +0.13
    0.6%
    511
    +0.09
    0.4%
    48
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    750
    +0.13
    0.03
    372
    +0.09
    0.03
    582
    +0.09
    0.01
    Negative Logits
    raft
    -0.75
    ateurs
    -0.74
    ORTS
    -0.74
     ®
    -0.73
    artisan
    -0.72
    inois
    -0.71
     priced
    -0.71
     inexpensive
    -0.70
     racket
    -0.70
    iberal
    -0.70
    POSITIVE LOGITS
     flashbacks
    1.35
     flashback
    1.28
     storyline
    1.28
     backstory
    1.27
    Characters
    1.16
     Characters
    1.13
     characters
    1.12
     protagonist
    1.06
     Whedon
    1.06
     storylines
    1.05
    Activations Density 0.627%

    No Known Activations