INDEX

Explanations

instances of capitalized words

oai_token-act-pair · gpt-3.5-turbo

references to "caps" or discussions about cap-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ĪĴ

-0.80

Ô

-0.76

¿½

-0.71

vironments

-0.68

Else

-0.68

 Answers

-0.67

ellig

-0.66

obj

-0.65

Eva

-0.65

ITNESS

-0.65

POSITIVE LOGITS

 caps

1.35

icum

1.11

ules

0.99

caps

0.97

 capped

0.97

mith

0.95

aic

0.94

itol

0.93

itals

0.82

 pots

0.79

Activations Density 0.004%

instances of capitalized words

references to "caps" or discussions about cap-related topics

No Comments

No Known Activations