INDEX

Explanations

text with characters and symbols not typically found in English

oai_token-act-pair · gpt-3.5-turbo

sequences of special characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 GOODMAN

-0.83

alos

-0.78

stro

-0.75

itar

-0.73

thumbnails

-0.70

rified

-0.69

stroke

-0.69

kid

-0.69

minster

-0.68

raviolet

-0.68

POSITIVE LOGITS

¬¼

1.25

Ĥ

1.09

į

1.08

¹

1.08

ħ

1.06

¬

1.05

¶ħ

1.03

Ķ

1.03

¯

1.03

²

1.01

Activations Density 0.006%

text with characters and symbols not typically found in English

sequences of special characters

No Comments

No Known Activations