© 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
    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. 13765
    Prev
    Next
    INDEX
    Explanations

    text related to setting up or following instructions for an account, medical procedures, or area-specific terms such as medical conditions, and common actions and terms related to game/app development and community engagement

    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
    447
    +0.12
    0.6%
    588
    +0.10
    0.5%
    258
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.28
    575
    +0.10
    0.20
    225
    +0.09
    0.12
    Negative Logits
    ãĥ£
    -0.72
    ãģį
    -0.71
    iana
    -0.71
    ãĥĥ
    -0.69
    ãĤ¨ãĥ«
    -0.67
    ificent
    -0.66
    bright
    -0.65
    amen
    -0.64
    angelo
    -0.63
    Gh
    -0.62
    POSITIVE LOGITS
     solution
    0.71
     solutions
    0.70
     (~
    0.69
     scenario
    0.64
     (<
    0.62
     scenarios
    0.61
     (>
    0.61
    edly
    0.60
     specialist
    0.58
     charm
    0.58
    Activations Density 8.936%

    No Known Activations