© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    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. 12971
    Prev
    Next
    INDEX
    Explanations

    expressions of desires or hopes

    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
    13
    +0.06
    0.5%
    16
    +0.06
    0.5%
    641
    +0.06
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    568
    +0.06
    0.05
    272
    +0.06
    0.08
    16
    +0.06
    0.07
    Negative Logits
    onym
    -0.45
     Flavoring
    -0.44
    itu
    -0.44
    orously
    -0.43
    acus
    -0.42
    olson
    -0.41
    helle
    -0.41
    Champ
    -0.41
    regular
    -0.41
    iltr
    -0.41
    POSITIVE LOGITS
    bone
    0.55
     fulfil
    0.54
    bones
    0.54
     fulfilled
    0.49
     wishes
    0.48
     Wish
    0.48
     fulfillment
    0.47
     wish
    0.45
    ign
    0.45
    reprene
    0.44
    Activations Density 10.226%

    No Known Activations