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. 45437
    Prev
    Next
    INDEX
    Explanations

    phrases related to time and frequency

    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
    +0.24
    1.2%
    430
    +0.11
    0.6%
    241
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    460
    +0.24
    0.10
    375
    +0.11
    0.05
    29
    +0.09
    0.13
    Negative Logits
    rats
    -0.63
    oir
    -0.61
    iar
    -0.60
    ston
    -0.60
    mie
    -0.58
    achev
    -0.56
    emetery
    -0.56
    Chan
    -0.55
    rights
    -0.55
    iat
    -0.55
    POSITIVE LOGITS
    hift
    0.79
    entimes
    0.73
     periods
    0.70
     moments
    0.66
     times
    0.65
     bouts
    0.64
    hops
    0.63
    cale
    0.62
     scenes
    0.62
    icum
    0.61
    Activations Density 9.984%

    No Known Activations