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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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 E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 798
    Prev
    Next
    INDEX
    Explanations

    short written pieces that provide information or updates on various topics, such as technology, sport events, or personal experiences

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-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
    152
    +0.16
    0.5%
    276
    +0.14
    0.5%
    182
    +0.12
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    85
    +0.16
    0.11
    101
    +0.14
    0.16
    231
    +0.12
    0.10
    Negative Logits
     respons
    -1.15
     thereafter
    -1.07
    è£
    -1.00
    ãĤ©
    -0.98
     thereof
    -0.97
     Vaugh
    -0.96
    itud
    -0.94
     versa
    -0.94
    âĶ
    -0.93
     afterwards
    -0.93
    POSITIVE LOGITS
     Expand
    1.29
    zbollah
    1.23
    chwitz
    1.11
    anmar
    1.08
    hesda
    1.06
     Caption
    1.04
    resa
    1.02
     Answer
    1.00
    reetings
    0.99
    lando
    0.98
    Activations Density 3.108%

    No Known Activations