Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 4569
    Prev
    Next
    INDEX
    Explanations

    phrases related to time, specifically mentioning a timeframe of "a couple of" something

    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
    68
    +0.03
    0.5%
    595
    +0.03
    0.5%
    716
    +0.03
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    68
    +0.03
    0.05
    702
    +0.03
    0.08
    509
    +0.03
    0.05
    Negative Logits
    ignant
    -0.19
    illin
    -0.18
    cont
    -0.18
     Schwar
    -0.17
     CARD
    -0.17
    heid
    -0.17
    etermined
    -0.16
     gall
    -0.16
     Delicious
    -0.16
     ABOUT
    -0.16
    POSITIVE LOGITS
    acles
    0.22
    ations
    0.21
    encies
    0.21
    ples
    0.21
    ricular
    0.21
    heed
    0.21
    liest
    0.21
    pora
    0.20
    ologies
    0.20
    pher
    0.20
    Activations Density 7.724%

    No Known Activations