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. 45251
    Prev
    Next
    INDEX
    Explanations

    functions and code snippets in a computer program

    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
    120
    +0.07
    0.5%
    440
    +0.06
    0.5%
    234
    +0.06
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    756
    +0.07
    -0.09
    393
    +0.06
    0.16
    159
    +0.06
    0.05
    Negative Logits
    aturdays
    -0.45
    aris
    -0.43
    abouts
    -0.43
    ayn
    -0.40
    udi
    -0.40
    agos
    -0.39
     congr
    -0.39
    ildo
    -0.39
    ALD
    -0.37
    othal
    -0.37
    POSITIVE LOGITS
    Parameter
    0.43
    ã쮿
    0.42
    ential
    0.42
    ifier
    0.41
     Mechdragon
    0.40
    ä¹ĭ
    0.40
    ãĤ»
    0.38
    ãĥĥãĥĪ
    0.38
    notation
    0.38
    ãĤ´
    0.38
    Activations Density 4.067%

    No Known Activations