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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 22655
    Prev
    Next
    INDEX
    Explanations

    references to personal pronouns and verb forms that indicate intentions or past actions

    oai_token-act-pair · gpt-4o-miniTriggered 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
    86
    +0.11
    0.5%
    532
    +0.10
    0.4%
    36
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.11
    0.07
    566
    +0.10
    0.05
    477
    +0.09
    0.04
    Negative Logits
    icle
    -0.61
     Fail
    -0.61
     obscurity
    -0.60
     Lit
    -0.60
    cial
    -0.59
     Chall
    -0.59
     Conquest
    -0.58
     Deadly
    -0.58
     Klu
    -0.57
     Karn
    -0.56
    POSITIVE LOGITS
     intend
    1.22
     intention
    1.20
     intends
    1.13
     regretted
    1.06
     intent
    1.03
     intentions
    1.02
     wished
    0.89
     wants
    0.89
     regrets
    0.87
     regret
    0.86
    Activations Density 0.287%

    No Known Activations