© 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. 32954
    Prev
    Next
    INDEX
    Explanations

    legal terminology and philosophical concepts

    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
    +0.28
    1.5%
    393
    +0.19
    1.0%
    369
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.28
    0.20
    546
    +0.19
    0.18
    100
    +0.10
    0.16
    Negative Logits
    ril
    -0.62
    ocial
    -0.58
    ï¸
    -0.56
    raft
    -0.56
    ction
    -0.55
    oen
    -0.53
    ulic
    -0.53
    rave
    -0.52
    othermal
    -0.52
    opal
    -0.51
    POSITIVE LOGITS
     etc
    1.33
    etc
    1.09
     respectively
    0.74
    whatever
    0.71
     blah
    0.67
     Flavoring
    0.61
    FontSize
    0.61
    artifacts
    0.59
     ect
    0.57
     rinse
    0.57
    Activations Density 9.245%

    No Known Activations