© 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. 6058
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical acts of hitting or striking

    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
    636
    +0.12
    0.5%
    326
    +0.12
    0.5%
    497
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.12
    0.09
    288
    +0.12
    0.10
    526
    +0.11
    -0.00
    Negative Logits
    ernand
    -0.79
    mpeg
    -0.77
    mental
    -0.76
    aghd
    -0.76
    gnu
    -0.74
    mingham
    -0.73
    rosso
    -0.70
    elsen
    -0.70
    ests
    -0.70
    éĸ
    -0.69
    POSITIVE LOGITS
     dab
    1.14
    creen
    0.97
     Splash
    0.89
    stick
    0.88
    dash
    0.86
     slapped
    0.83
    down
    0.83
     slap
    0.79
     goodbye
    0.77
    heet
    0.76
    Activations Density 1.249%

    No Known Activations