INDEX

Explanations

phrases relating to controversial or heated debates

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Cześć

-0.68

idać

-0.67

ഒ

-0.66

Samen

-0.63

Și

-0.60

Díky

-0.60

Czym

-0.59

 ettei

-0.59

Več

-0.59

ypeł

-0.59

POSITIVE LOGITS

tawan

0.64

nicity

0.61

 deleter

0.59

vanced

0.57

cyclopedia

0.56

 inappro

0.56

bership

0.56

 unve

0.56

 stratigraph

0.56

jango

0.55

Activations Density 0.561%

phrases relating to controversial or heated debates

No Comments

No Known Activations