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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 23841
    Prev
    Next
    INDEX
    Explanations

    references to felony charges in criminal contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    288
    +0.12
    0.6%
    90
    +0.09
    0.4%
    408
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    90
    +0.12
    0.01
    535
    +0.09
    0.01
    280
    +0.09
    0.01
    Negative Logits
    TPPStreamerBot
    -0.91
    arest
    -0.84
    abeth
    -0.82
    bits
    -0.82
    este
    -0.76
    arity
    -0.75
    tz
    -0.75
    obyl
    -0.74
    rx
    -0.74
    atives
    -0.73
    POSITIVE LOGITS
     convictions
    1.11
     punishable
    1.08
     conviction
    1.05
     probation
    1.02
     felony
    0.96
     charges
    0.96
     indictment
    0.94
     sentencing
    0.94
     felon
    0.93
     arrest
    0.93
    Activations Density 0.006%

    No Known Activations