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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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 L0 E2E Recon
    5. 2-RES_SLEFR-AJT
    6. 5426
    Prev
    Next
    INDEX
    Explanations

    instances of the word "compassion" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_slefr-ajt/2-res_slefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    4
    +0.06
    0.4%
    693
    +0.06
    0.4%
    131
    +0.05
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    626
    +0.06
    0.01
    331
    +0.06
    0.01
    147
    +0.05
    0.01
    Negative Logits
     aft
    -0.41
    ibli
    -0.40
    ideon
    -0.39
    package
    -0.39
    paralle
    -0.38
     psychiat
    -0.38
     underest
    -0.38
    minster
    -0.38
    avorite
    -0.37
    illard
    -0.37
    POSITIVE LOGITS
    assion
    0.47
     Zeal
    0.44
    aire
    0.44
     Corpus
    0.42
     Lug
    0.40
     Passion
    0.40
    anism
    0.39
    alg
    0.39
     moderation
    0.38
    itably
    0.38
    Activations Density 0.001%

    No Known Activations