© 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
    3. Joseph Bloom · Initial Gemma-2B-IT SAEs
    5. Gemma-2B-IT
    7. Residual Stream
    9. 12-RES-JB
    11. 3430
    Prev
    Next
    INDEX
    Explanations

    references to harassment, threats, and intimidation in online spaces

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    990
    +0.11
    0.4%
    1137
    +0.11
    0.3%
    1691
    +0.11
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1137
    +0.11
    0.04
    1691
    +0.11
    0.03
    944
    +0.11
    0.02
    Negative Logits
     spie
    -0.69
     poichè
    -0.68
     altrimenti
    -0.67
     perciò
    -0.65
     utop
    -0.64
     espri
    -0.63
     nemmeno
    -0.63
     impon
    -0.61
     gius
    -0.60
     desideri
    -0.60
    POSITIVE LOGITS
     harassment
    0.94
     withal
    0.93
     shewn
    0.91
     Shakspeare
    0.81
     harassing
    0.79
     Harassment
    0.78
     blest
    0.77
     harassed
    0.77
     tolerably
    0.75
     harass
    0.75
    Activations Density 0.169%

    No Known Activations