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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 32164
    Prev
    Next
    INDEX
    Explanations

    detailed instructions for building or assembling something, particularly using specific measurements and tools

    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
    +0.29
    1.4%
    420
    +0.17
    0.8%
    326
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.29
    0.11
    738
    +0.17
    0.08
    640
    +0.13
    0.35
    Negative Logits
    Social
    -0.89
    ocial
    -0.85
     Venezuel
    -0.84
     flashbacks
    -0.81
     famine
    -0.78
     Social
    -0.77
     Genocide
    -0.76
    Politics
    -0.74
     millennial
    -0.73
     Sana
    -0.73
    POSITIVE LOGITS
     clamp
    1.20
     diameter
    1.17
     connectors
    1.14
     adjustable
    1.12
     molded
    1.11
     weld
    1.10
     solder
    1.10
     thickness
    1.10
     removable
    1.09
     adhesive
    1.09
    Activations Density 28.939%

    No Known Activations