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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 21467
    Prev
    Next
    INDEX
    Explanations

    scientific and technical terms related to physics and astronomy

    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
    686
    +0.11
    0.5%
    654
    +0.11
    0.5%
    720
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.11
    0.04
    654
    +0.11
    0.03
    686
    +0.10
    0.03
    Negative Logits
    grand
    -0.88
    ortium
    -0.84
    john
    -0.78
    moon
    -0.77
    arij
    -0.77
    emetery
    -0.77
    widget
    -0.76
    arc
    -0.74
    Champ
    -0.74
    ammy
    -0.70
    POSITIVE LOGITS
     stimuli
    1.18
     adversity
    1.05
     surroundings
    0.84
     cues
    0.84
     hearing
    0.83
     oxygen
    0.83
     shocks
    0.80
     encountering
    0.80
     adapting
    0.80
     stimulus
    0.79
    Activations Density 0.171%

    No Known Activations