© 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
    3. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    5. GPT2-Small
    7. Residual Similar CE E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 5563
    Prev
    Next
    INDEX
    Explanations

    critiques and discussions of political and societal issues, specifically related to ideology, culture, and societal changes

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-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
    393
    +0.35
    1.3%
    36
    +0.20
    0.8%
    640
    +0.15
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.35
    0.30
    362
    +0.20
    0.17
    159
    +0.15
    0.15
    Negative Logits
    agents
    -0.97
    lest
    -0.93
    oons
    -0.92
    adding
    -0.91
    aunts
    -0.91
    Reviewed
    -0.87
    ettes
    -0.86
    evidence
    -0.84
    eworks
    -0.84
    bots
    -0.83
    POSITIVE LOGITS
     lot
    1.22
     serious
    1.20
     huge
    1.19
    usterity
    1.19
     massive
    1.18
     bunch
    1.17
     lengthy
    1.14
     relatively
    1.13
     considerable
    1.12
     gigantic
    1.11
    Activations Density 2.958%

    No Known Activations