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. 40189
    Prev
    Next
    INDEX
    Explanations

    instances of actions and consequences within various unfamiliar texts

    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
    +0.23
    1.3%
    379
    +0.09
    0.5%
    91
    +0.08
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.23
    0.18
    761
    +0.09
    0.11
    59
    +0.08
    0.09
    Negative Logits
    endif
    -0.62
    ciating
    -0.57
    fortunately
    -0.54
    depending
    -0.53
    assuming
    -0.53
    erm
    -0.52
    gradation
    -0.51
    Depending
    -0.50
    fter
    -0.49
    Air
    -0.49
    POSITIVE LOGITS
     illegal
    0.61
     sake
    0.60
     improper
    0.57
     purposes
    0.50
    tein
    0.50
     insensitive
    0.49
     nonviolent
    0.48
     wrongly
    0.48
    ãĤ¢ãĥ«
    0.48
     insufficient
    0.47
    Activations Density 11.308%

    No Known Activations