© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE Local
    5. 6-RES_SCL-AJT
    6. 35180
    Prev
    Next
    INDEX
    Explanations

    Proper nouns and names, particularly of people

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    497
    +0.09
    0.5%
    420
    +0.08
    0.4%
    756
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    159
    +0.09
    0.07
    730
    +0.08
    0.08
    326
    +0.08
    0.19
    Negative Logits
    orsi
    -0.60
    DERR
    -0.59
     Enhancement
    -0.57
    artney
    -0.54
    INESS
    -0.54
     highlights
    -0.52
    assets
    -0.52
     "$:/
    -0.50
     GPA
    -0.49
     highlight
    -0.49
    POSITIVE LOGITS
    ÃŁ
    0.87
    lla
    0.81
    zel
    0.79
    vre
    0.78
    bold
    0.78
    ppel
    0.77
    zing
    0.74
    zb
    0.74
    llo
    0.73
    hl
    0.72
    Activations Density 7.138%

    No Known Activations