INDEX

Explanations

phrases related to legal/judicial phrases and procedures

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

New Auto-Interp

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

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<bos>

-0.96

 близь

-0.64

Sa

-0.64

 використання

-0.62

Pa

-0.61

 умова

-0.60

 Alabama

-0.60

Fo

-0.60

Da

-0.60

 безпе

-0.59

POSITIVE LOGITS

 milano

1.57

 napoli

1.56

 imposs

1.55

 sentra

1.52

 abnorm

1.51

 bandung

1.48

 canel

1.46

 indestru

1.45

 disreg

1.43

 maneu

1.43

Activations Density 0.431%

phrases related to legal/judicial phrases and procedures

No Comments

No Known Activations