© 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 EvalsExportsAPI 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. 33751
    Prev
    Next
    INDEX
    Explanations

    text related to political matters, particularly discussions about democracy, investigations, political ideologies, and governmental actions

    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
    +0.27
    1.3%
    393
    +0.15
    0.8%
    665
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    138
    +0.27
    0.52
    447
    +0.15
    0.53
    64
    +0.14
    -0.14
    Negative Logits
    llor
    -0.67
    killed
    -0.64
     sacked
    -0.63
    awei
    -0.61
    wine
    -0.60
    CN
    -0.60
    interrupted
    -0.57
    HAEL
    -0.57
     shalt
    -0.57
     lodged
    -0.56
    POSITIVE LOGITS
    roads
    1.19
     regards
    1.19
    clusions
    1.16
     spite
    1.14
     accordance
    1.07
     lieu
    1.07
    efficiency
    1.05
    versions
    1.04
     relation
    1.03
     regard
    1.02
    Activations Density 6.963%

    No Known Activations