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. 13314
    Prev
    Next
    INDEX
    Explanations

    words related to crafting and sewing, as well as technical instructions and materials required for specific creative projects

    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
    +0.15
    0.6%
    527
    +0.12
    0.5%
    762
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    447
    +0.15
    0.45
    138
    +0.12
    0.43
    64
    +0.12
    -0.19
    Negative Logits
    patient
    -0.77
    noon
    -0.76
    ãĥį
    -0.76
    Operation
    -0.73
    York
    -0.71
    76561
    -0.71
    Installation
    -0.70
    Legend
    -0.68
    antha
    -0.68
    Vert
    -0.68
    POSITIVE LOGITS
     alike
    1.93
     respectively
    1.44
    hips
    1.33
     thereof
    1.33
    paces
    1.20
    hops
    1.13
    mith
    1.09
    uits
    1.08
    peed
    1.03
    hare
    1.02
    Activations Density 10.279%

    No Known Activations