© 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. 35281
    Prev
    Next
    INDEX
    Explanations

    sequences of words related to computer programming, system configuration, and mathematical operations

    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
    +0.07
    0.8%
    143
    +0.05
    0.5%
    533
    +0.04
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    143
    +0.07
    0.08
    642
    +0.05
    0.03
    738
    +0.04
    0.06
    Negative Logits
    twitch
    -0.33
    Story
    -0.32
    girls
    -0.32
    president
    -0.32
    cffffcc
    -0.31
    iste
    -0.31
    tumblr
    -0.30
    flame
    -0.30
    competitive
    -0.29
    itters
    -0.29
    POSITIVE LOGITS
     duplicate
    0.39
     unnecessary
    0.37
     parentheses
    0.36
     existing
    0.35
     brackets
    0.34
     unused
    0.34
     ~/.
    0.34
     initialization
    0.33
     redundant
    0.33
     subscript
    0.32
    Activations Density 8.654%

    No Known Activations