© 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 Local
    9. 6-RES_SCL-AJT
    11. 40483
    Prev
    Next
    INDEX
    Explanations

    references to events, places, and people

    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
    343
    +0.09
    0.5%
    59
    +0.09
    0.5%
    718
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.09
    0.21
    288
    +0.09
    0.21
    151
    +0.08
    0.15
    Negative Logits
    onom
    -0.66
    olson
    -0.59
    onymous
    -0.58
    lass
    -0.58
    umper
    -0.56
    ocalypse
    -0.56
    eering
    -0.55
    erity
    -0.54
     detach
    -0.54
    hog
    -0.54
    POSITIVE LOGITS
    Sweet
    0.70
    ven
    0.69
     Mae
    0.69
     Bee
    0.69
    iva
    0.66
    uthor
    0.64
    BLE
    0.63
    Berry
    0.63
    VE
    0.63
    cess
    0.62
    Activations Density 6.701%

    No Known Activations