© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 5509
    Prev
    Next
    INDEX
    Explanations

    instances where there is an assertion that something is incorrect or mistaken

    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
    547
    +0.06
    0.5%
    608
    +0.06
    0.5%
    738
    +0.05
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    487
    +0.06
    0.03
    278
    +0.06
    0.03
    568
    +0.05
    0.02
    Negative Logits
    urable
    -0.42
    ochet
    -0.39
    mens
    -0.39
    dom
    -0.39
    DIT
    -0.38
    ensable
    -0.37
    orneys
    -0.37
    hes
    -0.37
    faced
    -0.35
    hend
    -0.35
    POSITIVE LOGITS
    othal
    0.43
     misplaced
    0.41
    >[
    0.40
     variety
    0.39
     guiActiveUn
    0.38
    ij士
    0.36
     stereotype
    0.35
     Lone
    0.35
     disappro
    0.35
     Shuttle
    0.33
    Activations Density 0.231%

    No Known Activations