© 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
    3. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    5. GPT2-Small
    7. Residual Similar CE Local
    9. 6-RES_SCL-AJT
    11. 3973
    Prev
    Next
    INDEX
    Explanations

    information about ongoing investigations or developments in a given situation, including details on contacting authorities for tips or assistance

    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
    447
    +0.20
    0.9%
    393
    +0.16
    0.7%
    287
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    138
    +0.20
    0.33
    447
    +0.16
    0.34
    64
    +0.12
    -0.13
    Negative Logits
    full
    -0.66
     Khe
    -0.61
     CJ
    -0.59
    crow
    -0.57
    Bio
    -0.56
     che
    -0.55
     Mecca
    -0.55
     Charleston
    -0.55
     Magicka
    -0.55
    ointed
    -0.54
    POSITIVE LOGITS
    abouts
    1.35
    upon
    1.00
    fore
    0.88
    FORE
    0.77
    oplan
    0.70
     aren
    0.69
    olver
    0.69
     are
    0.68
    're
    0.66
     weren
    0.66
    Activations Density 6.265%

    No Known Activations