© 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. 40013
    Prev
    Next
    INDEX
    Explanations

    instances of the word "expanding"

    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
    169
    +0.10
    0.5%
    208
    +0.10
    0.5%
    214
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.17
    159
    +0.10
    0.06
    355
    +0.09
    0.05
    Negative Logits
    ball
    -0.81
    âĸ¬âĸ¬
    -0.73
    footed
    -0.71
    warts
    -0.70
    ters
    -0.70
    balls
    -0.68
    ãĥ£
    -0.67
    milo
    -0.64
    tering
    -0.61
    creen
    -0.61
    POSITIVE LOGITS
    onential
    1.33
    ansion
    1.25
    onent
    1.18
    osed
    1.18
    anded
    1.18
    ulsion
    1.12
    ression
    1.10
    anding
    1.02
    ressive
    1.00
    ressed
    0.97
    Activations Density 6.615%

    No Known Activations