© 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. 16873
    Prev
    Next
    INDEX
    Explanations

    topics of discussion or debate

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    195
    +0.16
    0.7%
    496
    +0.12
    0.5%
    336
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    195
    +0.16
    0.03
    336
    +0.12
    0.03
    163
    +0.11
    0.03
    Negative Logits
    ramid
    -0.83
    urses
    -0.83
    ellow
    -0.76
    glas
    -0.76
    ãĥ³ãĤ¸
    -0.76
    ramids
    -0.76
    ongyang
    -0.74
    emouth
    -0.74
    ravis
    -0.72
    mingham
    -0.72
    POSITIVE LOGITS
     plag
    1.04
     confronting
    0.96
     arising
    0.89
     facing
    0.88
     affecting
    0.88
     relating
    0.86
     flared
    0.84
     raised
    0.84
     unresolved
    0.83
     resolved
    0.81
    Activations Density 0.054%

    No Known Activations