© 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 12288
    9. 8-RES_FS12288-JB
    11. 825
    Prev
    Next
    INDEX
    Explanations

    discussions about problems and difficulties

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    756
    +0.16
    0.7%
    259
    +0.11
    0.5%
    458
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    756
    +0.16
    -0.01
    102
    +0.11
    0.10
    718
    +0.10
    0.08
    Negative Logits
     slots
    -0.80
     nods
    -0.78
     choir
    -0.72
     chapel
    -0.72
    akeru
    -0.67
    urai
    -0.67
     ceremonies
    -0.67
    ACP
    -0.66
     promo
    -0.66
     loudspe
    -0.65
    POSITIVE LOGITS
     exacerbated
    1.34
     plag
    1.26
     worsened
    1.21
     caused
    1.21
     severity
    1.16
     endemic
    1.16
     inflicted
    1.12
     plague
    1.10
     aggravated
    1.08
     compounded
    1.06
    Activations Density 1.189%

    No Known Activations