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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & 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. 45571
    Prev
    Next
    INDEX
    Explanations

    profanity and aggressive language

    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
    104
    +0.12
    0.6%
    213
    +0.11
    0.5%
    637
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    536
    +0.12
    0.05
    372
    +0.11
    0.06
    620
    +0.10
    0.01
    Negative Logits
    ISC
    -0.99
    Newsletter
    -0.94
    Interstitial
    -0.94
    Reviewer
    -0.84
    xit
    -0.82
    krit
    -0.81
    031
    -0.81
    MSN
    -0.79
    quished
    -0.75
    RANT
    -0.75
    POSITIVE LOGITS
     wanna
    0.76
     dop
    0.76
     tha
    0.75
     lifetime
    0.73
     dat
    0.71
     punk
    0.68
    pace
    0.65
     queer
    0.64
     kid
    0.64
     gonna
    0.63
    Activations Density 5.376%

    No Known Activations