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 · Initial Gemma-2B SAEs
    3. Gemma-2B
    4. Residual Stream
    5. 6-RES-JB
    6. 654
    Prev
    Next
    INDEX
    Explanations

    violent actions and intense emotions, particularly related to screaming

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.6.hook_resid_post_16384_anthropic_fast_lr
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.15
    0.7%
    321
    +0.11
    0.5%
    1416
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    321
    +0.15
    0.04
    1490
    +0.11
    0.03
    1363
    +0.11
    0.03
    Negative Logits
    <bos>
    -2.83
    lateinit
    -0.81
     nawr
    -0.77
    Esperamos
    -0.76
    ⓧ
    -0.74
    <?
    -0.74
    UnitTesting
    -0.72
    Galería
    -0.72
    util
    -0.71
    HasIndex
    -0.71
    POSITIVE LOGITS
     ecru
    2.17
     swarovski
    2.13
     impractica
    2.13
     indestru
    2.11
     increa
    2.09
     unlaw
    2.00
     impra
    2.00
     unden
    1.98
     reluct
    1.93
     disreg
    1.93
    Activations Density 0.191%

    No Known Activations