Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 36
    Prev
    Next
    INDEX
    Explanations

    information related to crime and tragic events

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    480
    +0.12
    0.5%
    15
    +0.11
    0.5%
    361
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    576
    +0.12
    0.07
    561
    +0.11
    0.15
    15
    +0.11
    0.17
    Negative Logits
    Newsletter
    -1.02
     etc
    -0.78
    Secondly
    -0.75
    =-
    -0.71
     Secondly
    -0.69
    âĨij
    -0.68
    )/
    -0.68
    /-
    -0.68
    peak
    -0.66
    CONCLUS
    -0.65
    POSITIVE LOGITS
     embattled
    1.03
     Wednesday
    0.93
     Thursday
    0.91
     Tuesday
    0.88
     unveiled
    0.86
     TODAY
    0.85
     Monday
    0.84
     watchdog
    0.83
     Friday
    0.83
     bombshell
    0.82
    Activations Density 9.964%

    No Known Activations