© 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. 14079
    Prev
    Next
    INDEX
    Explanations

    references to software-related terms

    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
    704
    +0.10
    0.5%
    726
    +0.10
    0.5%
    148
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    137
    +0.10
    0.01
    723
    +0.10
    0.01
    658
    +0.08
    0.01
    Negative Logits
    leasing
    -0.72
    igne
    -0.71
    ipment
    -0.71
    OIL
    -0.71
    stals
    -0.69
    outer
    -0.67
    bell
    -0.67
    rities
    -0.66
    shaw
    -0.65
    osc
    -0.65
    POSITIVE LOGITS
     Shepherd
    0.70
     Chimera
    0.60
     Dwarf
    0.59
     remembrance
    0.59
     CoC
    0.56
     Summit
    0.55
     Pavilion
    0.55
     Pathfinder
    0.55
     Butterfly
    0.55
     prevail
    0.54
    Activations Density 0.012%

    No Known Activations