INDEX

Explanations

out words related to researching or exploring a topic in depth

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

 emphat

-1.25

fte

-1.10

 effe

-1.08

 wien

-1.08

fta

-1.06

 intermitt

-1.04

 affor

-1.03

 reluct

-1.03

 perfet

-1.03

 accla

-1.01

POSITIVE LOGITS

 learn

0.85

<bos>

0.77

learn

0.75

Learn

0.70

 Learn

0.68

 learns

0.68

 learned

0.66

how

0.66

 discover

0.66

out

0.66

Activations Density 0.055%

out words related to researching or exploring a topic in depth

No Comments

No Known Activations