© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 11738
    Prev
    Next
    INDEX
    Explanations

    pronouns referring to specific individuals or groups, particularly when the referring individual or group is the subject of a sentence

    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
    151
    +0.14
    0.5%
    371
    +0.12
    0.4%
    93
    +0.12
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    553
    +0.14
    0.04
    188
    +0.12
    0.04
    477
    +0.12
    0.03
    Negative Logits
    onal
    -1.24
    olid
    -1.18
    ammy
    -1.16
     RELEASE
    -1.09
    eway
    -1.05
    Dispatch
    -1.00
    veyard
    -1.00
    =-=-=-=-=-=-=-=-
    -0.99
    oxin
    -0.98
    yip
    -0.97
    POSITIVE LOGITS
     profess
    1.10
     christ
    1.04
     admitted
    1.03
     validated
    1.00
     confessed
    0.99
     personally
    0.99
     belonged
    0.97
    é¾įåĸļ士
    0.97
     admits
    0.97
     knew
    0.96
    Activations Density 0.664%

    No Known Activations