© 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. 1186
    Prev
    Next
    INDEX
    Explanations

    information related to criminal investigations and legal proceedings

    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
    352
    +0.09
    0.4%
    5
    +0.09
    0.4%
    718
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    718
    +0.09
    0.06
    439
    +0.09
    0.05
    591
    +0.08
    0.04
    Negative Logits
    ortment
    -0.77
    itton
    -0.76
    yssey
    -0.67
    sen
    -0.66
     [|
    -0.65
    igion
    -0.63
    ahime
    -0.62
    kk
    -0.61
    eals
    -0.61
     summers
    -0.60
    POSITIVE LOGITS
     whatsoever
    1.52
     nor
    1.31
     anymore
    1.06
     anywhere
    0.98
     except
    0.96
    nor
    0.94
    except
    0.86
     anybody
    0.84
     conclusive
    0.83
     satisfactory
    0.82
    Activations Density 0.568%

    No Known Activations