Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 1087
    Prev
    Next
    INDEX
    Explanations

    a specific pattern of characters that might be related to a particular language or script

    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
    491
    +0.09
    0.5%
    626
    +0.08
    0.5%
    759
    +0.06
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    107
    +0.09
    0.02
    291
    +0.08
    0.02
    284
    +0.06
    0.02
    Negative Logits
    zees
    -0.69
    swick
    -0.62
    zee
    -0.61
    abase
    -0.61
    ormal
    -0.57
    ittens
    -0.56
    sole
    -0.56
    ilateral
    -0.56
    ongs
    -0.56
    rish
    -0.55
    POSITIVE LOGITS
    Ö
    0.77
    Ö¼
    0.75
    ÙIJ
    0.70
    ×Ļ
    0.67
    ×IJ
    0.64
    ×Ķ
    0.62
    Ù
    0.61
    Ùĩ
    0.61
    ×ķ
    0.60
    ÙĬ
    0.59
    Activations Density 0.070%

    No Known Activations