INDEX

Explanations

references to things being oriented or positioned in an upside-down manner

oai_token-act-pair · gpt-3.5-turbo

the concept of "upside" or positive potential in various contexts

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

naire

-0.85

ocally

-0.78

nance

-0.74

ocal

-0.74

ãģ®å®

-0.73

ainers

-0.72

éļ

-0.72

Rush

-0.71

ions

-0.70

%"

-0.68

POSITIVE LOGITS

 upside

1.18

ï¸

0.79

ward

0.74

vable

0.72

zzle

0.71

verted

0.70

 infinity

0.69

 srfAttach

0.69

 blind

0.69

side

0.67

Activations Density 0.009%

references to things being oriented or positioned in an upside-down manner

the concept of "upside" or positive potential in various contexts

No Comments

No Known Activations