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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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 Local
    5. 6-RES_SCL-AJT
    6. 13307
    Prev
    Next
    INDEX
    Explanations

    instances of the word "blink" and related terms such as "instant" and "eye blink."

    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
    578
    +0.07
    0.7%
    471
    +0.05
    0.5%
    453
    +0.04
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.07
    0.22
    288
    +0.05
    0.22
    151
    +0.04
    0.18
    Negative Logits
    atro
    -0.35
    ilitarian
    -0.35
    ript
    -0.32
    %"
    -0.32
    ulative
    -0.32
    ilitary
    -0.32
    neau
    -0.32
    ivil
    -0.32
     possessions
    -0.32
    ÃįÃį
    -0.31
    POSITIVE LOGITS
     blink
    0.52
     blinking
    0.43
    ingly
    0.40
     blinked
    0.39
     wink
    0.38
    ĸļ
    0.38
     Eyes
    0.35
     aloud
    0.33
    eyes
    0.32
     twitch
    0.32
    Activations Density 7.235%

    No Known Activations