INDEX

Explanations

phrases related to being criticized or facing scrutiny

oai_token-act-pair · gpt-3.5-turbo

expressions related to being criticized or endangered

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

erd

-0.63

 depress

-0.63

equ

-0.63

alone

-0.59

ourage

-0.58

olen

-0.56

 Camden

-0.56

Lot

-0.55

annis

-0.55

 Loaded

-0.55

POSITIVE LOGITS

 scrutiny

0.81

ÃĽ

0.72

BAT

0.69

zai

0.69

Ø©

0.68

é¾įå¥ĳå£«

0.67

èª

0.67

å£«

0.65

Ú

0.64

ÅŁ

0.63

Activations Density 0.068%

phrases related to being criticized or facing scrutiny

expressions related to being criticized or endangered

No Comments

No Known Activations