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

    Neuronpedia

    APICircuit 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. 13596
    Prev
    Next
    INDEX
    Explanations

    scientific and technical terms related to functional analysis and structural modeling

    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
    +0.21
    1.0%
    447
    +0.15
    0.7%
    137
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.21
    0.19
    137
    +0.15
    0.04
    707
    +0.12
    0.07
    Negative Logits
    yang
    -0.69
    yard
    -0.66
     Preferred
    -0.65
    yah
    -0.65
    Tube
    -0.65
    gars
    -0.64
     Eck
    -0.64
    yards
    -0.63
     plaint
    -0.63
     Furious
    -0.62
    POSITIVE LOGITS
    istic
    1.04
    ities
    1.03
    istically
    0.94
    tarian
    0.94
    ized
    0.90
    tarians
    0.90
    acies
    0.87
    ization
    0.84
    ists
    0.83
    ative
    0.83
    Activations Density 5.313%

    No Known Activations