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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 34718
    Prev
    Next
    INDEX
    Explanations

    information related to technical implementation details and programming concepts

    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
    +0.09
    0.5%
    157
    +0.08
    0.4%
    48
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    675
    +0.09
    0.23
    338
    +0.08
    0.20
    48
    +0.08
    0.20
    Negative Logits
     mathemat
    -0.87
     disadvant
    -0.78
    ©¶æ¥µ
    -0.73
     veter
    -0.69
    enegger
    -0.67
     filib
    -0.65
     unbeliev
    -0.65
     vulner
    -0.64
     unnecess
    -0.64
     accomp
    -0.63
    POSITIVE LOGITS
    ï¸ı
    0.97
    ship
    0.82
    sand
    0.76
    kay
    0.68
    cation
    0.67
    s
    0.66
    sing
    0.66
    sure
    0.64
    Bah
    0.64
    forth
    0.64
    Activations Density 5.371%

    No Known Activations