© 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. 39921
    Prev
    Next
    INDEX
    Explanations

    opinions and viewpoints from different individuals on a specific topic or event

    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
    +0.05
    0.6%
    556
    +0.05
    0.5%
    691
    +0.04
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    393
    +0.05
    0.29
    745
    +0.05
    -0.17
    111
    +0.04
    0.09
    Negative Logits
     Thieves
    -0.33
     Andersen
    -0.30
     Lis
    -0.29
     Dull
    -0.29
    RIP
    -0.28
     virus
    -0.28
    adobe
    -0.27
     ILCS
    -0.27
     goose
    -0.27
     Marse
    -0.27
    POSITIVE LOGITS
    odox
    0.47
    igious
    0.37
     approving
    0.35
    tical
    0.35
    omo
    0.35
     PsyNet
    0.34
     interven
    0.33
    apse
    0.32
    onymous
    0.31
     front
    0.29
    Activations Density 13.175%

    No Known Activations