INDEX

Explanations

specific phrases indicating extreme or impactful situations

oai_token-act-pair · gpt-3.5-turbo

phrases expressing extreme consequences or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ramid

-0.76

usk

-0.73

sson

-0.73

zman

-0.71

ilver

-0.70

byss

-0.70

ãĤ©

-0.69

 redes

-0.69

intosh

-0.68

rolet

-0.68

POSITIVE LOGITS

 even

1.10

it

1.05

 nobody

1.03

 they

0.95

 anyone

0.91

 anybody

0.88

 none

0.83

 hardly

0.82

 whenever

0.81

we

0.80

Activations Density 0.135%

specific phrases indicating extreme or impactful situations

phrases expressing extreme consequences or situations

No Comments

No Known Activations