INDEX

Explanations

emotion-related words or phrases, especially negative emotions like anger, greed, frustration, and worry

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

 ?...

-1.28

 !...

-1.25

 emphat

-1.16

fuf

-1.11

 Præ

-1.09

 depic

-1.08

 embra

-1.07

aen

-1.06

 pessi

-1.05

 Godt

-1.03

POSITIVE LOGITS

<bos>

0.73

principalColumn

0.62

AddHtmlAttribute

0.56

 during

0.56

 again

0.56

 after

0.56

 quickly

0.54

 with

0.54

 Gets

0.53

ніципа

0.53

Activations Density 0.228%

emotion-related words or phrases, especially negative emotions like anger, greed, frustration, and worry

No Comments

No Known Activations