INDEX

Explanations

safety-related terms or concepts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

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

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<bos>

-3.32

/**

-0.81

<?

-0.80

ⓧ

-0.77

 rehabilitate

-0.66

 realize

-0.65

 reunite

-0.65

 exert

-0.63

 accumulate

-0.63

POSITIVE LOGITS

cæ

1.46

 socie

1.42

 magis

1.42

 Bibl

1.40

 wien

1.39

 bayern

1.38

 milano

1.38

 tramont

1.37

 nomine

1.36

 maroc

1.34

Activations Density 0.067%

safety-related terms or concepts

No Comments

No Known Activations