© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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 Local
    9. 6-RES_SCL-AJT
    11. 36932
    Prev
    Next
    INDEX
    Explanations

    places, locations, and geographical references, particularly related to the Baltic Sea and surrounding regions

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-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
    +0.09
    0.7%
    701
    +0.06
    0.5%
    99
    +0.06
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    48
    +0.09
    0.14
    372
    +0.06
    0.12
    675
    +0.06
    0.14
    Negative Logits
    aping
    -0.52
    aped
    -0.52
    perse
    -0.47
    oeuv
    -0.47
    IRED
    -0.46
    orses
    -0.44
    erving
    -0.44
    inct
    -0.44
    ovember
    -0.43
    ENTS
    -0.43
    POSITIVE LOGITS
     Tid
    0.50
     Baltic
    0.49
     States
    0.48
     Countries
    0.46
     Stream
    0.45
     Regions
    0.44
    Balt
    0.43
     Pry
    0.42
     Rivals
    0.42
     Bastard
    0.42
    Activations Density 10.151%

    No Known Activations