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. 440
    Prev
    Next
    INDEX
    Explanations

    questions or scenarios related to decision-making and responsibility

    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
    599
    +0.14
    0.4%
    638
    +0.14
    0.4%
    393
    +0.13
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    287
    +0.14
    0.04
    3
    +0.14
    0.04
    551
    +0.13
    0.04
    Negative Logits
    cule
    -0.98
    ahime
    -0.92
    iard
    -0.88
    lication
    -0.88
    ulhu
    -0.87
    iza
    -0.87
    zeb
    -0.86
    Lago
    -0.85
    fort
    -0.85
    pha
    -0.84
    POSITIVE LOGITS
     happen
    1.37
     happens
    1.16
     happened
    1.14
     transpired
    1.13
    ?]
    1.12
     happ
    1.06
     characterize
    1.01
     differe
    0.94
     difference
    0.93
     spoil
    0.91
    Activations Density 0.362%

    No Known Activations