INDEX

Explanations

phrases related to criticizing or discussing a particular subject

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

 wien

-1.63

 lidl

-1.51

 effe

-1.51

 depic

-1.50

aen

-1.50

 indestru

-1.48

 embra

-1.48

 suscep

-1.46

 waer

-1.46

 aquarelle

-1.45

POSITIVE LOGITS

 this

0.86

 particular

0.84

 thing

0.82

 THIS

0.77

 stuff

0.76

this

0.75

 whole

0.73

 kind

0.73

 year

0.73

 This

0.72

Activations Density 0.240%

phrases related to criticizing or discussing a particular subject

No Comments

No Known Activations