INDEX

Explanations

phrases indicating something being created or produced

oai_token-act-pair · gpt-3.5-turbo

phrases related to the impact or effectiveness of actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 withdrawing

-0.71

 adopting

-0.67

PN

-0.66

 relying

-0.66

 overseen

-0.64

Mand

-0.63

 openly

-0.62

 unilateral

-0.61

 Title

-0.60

 corpor

-0.60

POSITIVE LOGITS

 havoc

0.78

olor

0.76

 intrigue

0.75

flies

0.72

orage

0.70

rouse

0.69

vae

0.69

 sear

0.68

rome

0.66

wered

0.65

Activations Density 0.449%

phrases indicating something being created or produced

phrases related to the impact or effectiveness of actions

No Comments

No Known Activations