INDEX

Explanations

words related to emotional distress or agitation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "upset" and variations related to emotional distress

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

 livest

-0.96

glas

-0.91

 liner

-0.80

vernment

-0.77

istered

-0.76

audi

-0.74

icrobial

-0.74

hent

-0.73

zag

-0.72

atures

-0.72

POSITIVE LOGITS

dy

0.90

 upset

0.77

der

0.75

Brexit

0.71

die

0.69

 Voldemort

0.68

 Dumbledore

0.68

wart

0.66

 Diana

0.66

bur

0.64

Activations Density 0.011%

words related to emotional distress or agitation

instances of the word "upset" and variations related to emotional distress

No Comments

No Known Activations