© 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 E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 14059
    Prev
    Next
    INDEX
    Explanations

    words related to torture and physical abuse

    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
    288
    +0.20
    0.7%
    343
    +0.13
    0.5%
    363
    +0.12
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.20
    0.09
    343
    +0.13
    0.05
    326
    +0.12
    0.06
    Negative Logits
     donor
    -0.97
    earance
    -0.82
    BY
    -0.79
     donation
    -0.78
     iod
    -0.77
    ership
    -0.75
     Prospect
    -0.75
     magnification
    -0.74
    Ü
    -0.73
    ACP
    -0.72
    POSITIVE LOGITS
    urous
    1.37
    oise
    1.35
    illas
    1.11
    anamo
    0.97
    ured
    0.96
    aste
    0.95
    teenth
    0.95
    uous
    0.94
    uses
    0.92
    ificate
    0.92
    Activations Density 0.796%

    No Known Activations