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. 40285
    Prev
    Next
    INDEX
    Explanations

    terms related to email newsletter sign-ups and verification processes

    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
    527
    +0.13
    0.5%
    170
    +0.11
    0.5%
    355
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.13
    0.26
    393
    +0.11
    0.24
    745
    +0.10
    -0.15
    Negative Logits
    vironments
    -0.74
    fman
    -0.74
    cellaneous
    -0.70
    urat
    -0.70
    imer
    -0.70
    ween
    -0.69
    rans
    -0.68
    itive
    -0.68
    ebted
    -0.68
    orah
    -0.67
    POSITIVE LOGITS
    taboola
    0.98
     Invalid
    0.81
     anymore
    0.77
    Sorry
    0.76
     ISI
    0.74
     nor
    0.73
     invalid
    0.72
    dayName
    0.71
     Vaugh
    0.70
     Cao
    0.70
    Activations Density 8.878%

    No Known Activations