© 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. 13700
    Prev
    Next
    INDEX
    Explanations

    instances of the word "ensure" followed by an action or outcome

    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
    +0.25
    1.6%
    393
    +0.18
    1.2%
    481
    +0.09
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    679
    +0.25
    0.30
    393
    +0.18
    0.31
    283
    +0.09
    0.06
    Negative Logits
    ãĤ´ãĥ³
    -0.50
    HT
    -0.46
    gur
    -0.44
    orah
    -0.44
    aukee
    -0.44
    ET
    -0.42
     )]
    -0.40
    JM
    -0.40
    ãĤ¨ãĥ«
    -0.39
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    -0.39
    POSITIVE LOGITS
     they
    0.58
    soever
    0.54
     contradicts
    0.47
     there
    0.45
     we
    0.45
     existed
    0.45
     someone
    0.44
     occurred
    0.44
     happened
    0.43
     coma
    0.43
    Activations Density 14.052%

    No Known Activations