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 3072
    5. 8-RES_FS3072-JB
    6. 2483
    Prev
    Next
    INDEX
    Explanations

    words related to contracts or agreements

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    710
    +0.11
    0.5%
    277
    +0.11
    0.5%
    213
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.08
    140
    +0.11
    0.06
    76
    +0.10
    0.06
    Negative Logits
    nen
    -0.89
    CVE
    -0.83
    cffffcc
    -0.77
    Sax
    -0.70
     ambassadors
    -0.69
    obal
    -0.69
    rian
    -0.68
    ONSORED
    -0.67
    «ĺ
    -0.66
    Enlarge
    -0.65
    POSITIVE LOGITS
    clip
    1.24
     towels
    1.06
    clips
    1.04
     pus
    0.94
     towel
    0.86
     ink
    0.86
    pee
    0.84
    thin
    0.79
    issors
    0.79
     cartridges
    0.78
    Activations Density 0.889%

    No Known Activations