INDEX

Explanations

words related to accuracy or precision

oai_token-act-pair · gpt-3.5-turbo

references to acknowledgment or acceptance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

VID

-0.90

vous

-0.84

prototype

-0.81

geist

-0.79

 Dare

-0.79

çīĪ

-0.75

ä

-0.71

 hope

-0.71

hner

-0.70

éĹĺ

-0.69

POSITIVE LOGITS

reditation

1.26

redited

1.16

ursed

1.02

retion

0.99

enture

0.98

imate

0.92

ords

0.90

redit

0.89

acc

0.89

atech

0.89

Activations Density 0.007%

words related to accuracy or precision

references to acknowledgment or acceptance

No Comments

No Known Activations