© 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. 39971
    Prev
    Next
    INDEX
    Explanations

    technical terms and explanations related to a specific subject, likely in a written format

    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
    683
    +0.10
    0.6%
    137
    +0.09
    0.6%
    366
    +0.08
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    137
    +0.10
    0.06
    212
    +0.09
    0.08
    687
    +0.08
    0.09
    Negative Logits
    enance
    -0.64
    ourses
    -0.62
    Newsletter
    -0.56
    levision
    -0.53
    fund
    -0.52
    Govern
    -0.52
    antle
    -0.51
    ablish
    -0.50
     Superintendent
    -0.49
    ablished
    -0.49
    POSITIVE LOGITS
     kidding
    0.78
     haha
    0.77
     typo
    0.75
     guessed
    0.70
     forgot
    0.69
     spoiler
    0.69
    !).
    0.69
     kinda
    0.67
     ya
    0.67
     ;)
    0.67
    Activations Density 15.860%

    No Known Activations