© 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. 9890
    Prev
    Next
    INDEX
    Explanations

    words related to consulting and advisory services

    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
    679
    +0.13
    0.6%
    288
    +0.12
    0.5%
    726
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    679
    +0.13
    0.03
    88
    +0.12
    0.02
    288
    +0.10
    0.03
    Negative Logits
    bers
    -0.72
    corn
    -0.70
    eat
    -0.69
    stop
    -0.68
    Æ
    -0.67
    boats
    -0.64
    amy
    -0.63
    ending
    -0.63
    zees
    -0.63
    nings
    -0.62
    POSITIVE LOGITS
     consulted
    0.99
     advising
    0.93
     advise
    0.88
     consult
    0.81
     Consult
    0.80
     Pric
    0.77
    ancies
    0.77
     consultants
    0.76
     advisors
    0.76
    oga
    0.76
    Activations Density 0.089%

    No Known Activations