© 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. 2662
    Prev
    Next
    INDEX
    Explanations

    text related to problem-solving and decision-making processes

    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
    487
    +0.11
    0.5%
    313
    +0.09
    0.4%
    45
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    170
    +0.11
    0.03
    561
    +0.09
    0.03
    598
    +0.08
    0.02
    Negative Logits
    apo
    -0.63
    role
    -0.61
     targ
    -0.61
    Lago
    -0.59
    uggle
    -0.58
    aran
    -0.56
     fixme
    -0.56
    photos
    -0.56
    utton
    -0.56
    gard
    -0.56
    POSITIVE LOGITS
     finally
    1.36
     settled
    0.98
     conclude
    0.91
     recomm
    0.88
     culminating
    0.88
     decided
    0.88
     FIN
    0.81
     eventually
    0.81
     recons
    0.81
     settle
    0.81
    Activations Density 0.255%

    No Known Activations