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 E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 492
    Prev
    Next
    INDEX
    Explanations

    mentions of specific thoughts or ideas coming to someone's mind

    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
    65
    +0.03
    0.5%
    756
    +0.03
    0.5%
    753
    +0.02
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    753
    +0.03
    0.02
    386
    +0.03
    0.02
    125
    +0.02
    0.02
    Negative Logits
     resil
    -0.17
    ailability
    -0.16
     correctness
    -0.16
     divergence
    -0.15
    ifiable
    -0.15
     determination
    -0.15
     reliability
    -0.15
    idential
    -0.15
     confidentiality
    -0.15
     Consumer
    -0.15
    POSITIVE LOGITS
    mire
    0.17
    urch
    0.17
    ebted
    0.16
    ovie
    0.16
    ulz
    0.16
    ue
    0.15
    laugh
    0.15
    empl
    0.15
    merce
    0.15
    Cra
    0.15
    Activations Density 0.061%

    No Known Activations