© 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. 4083
    Prev
    Next
    INDEX
    Explanations

    technical terms related to software development and programming

    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
    481
    +0.24
    1.5%
    447
    +0.14
    0.8%
    393
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.24
    0.35
    255
    +0.14
    0.19
    6
    +0.11
    0.21
    Negative Logits
    unic
    -0.56
    iant
    -0.53
    leground
    -0.52
    Pacific
    -0.51
    ante
    -0.49
     populism
    -0.49
    reprene
    -0.47
    anya
    -0.47
    venture
    -0.47
    anto
    -0.46
    POSITIVE LOGITS
     itself
    0.48
    thus
    0.48
     convey
    0.46
     Hammond
    0.45
     whence
    0.44
     sink
    0.43
    Whereas
    0.43
     upd
    0.43
     BJ
    0.41
     bis
    0.41
    Activations Density 13.153%

    No Known Activations