INDEX

Explanations

the color "gray" followed by nouns or descriptive terms associated with diverse contexts

oai_token-act-pair · gpt-3.5-turbo

references to the color gray

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

=-=-=-=-

-0.99

========

-0.83

ÄŁ

-0.80

etics

-0.78

Technical

-0.73

Ì¶

-0.73

Witness

-0.73

itu

-0.72

ovie

-0.72

idem

-0.71

POSITIVE LOGITS

 gray

1.15

 grey

1.08

hound

1.03

gray

0.97

 whale

0.92

beard

0.89

 wolf

0.88

 shading

0.86

pup

0.85

 blur

0.85

Activations Density 0.008%

the color "gray" followed by nouns or descriptive terms associated with diverse contexts

references to the color gray

No Comments

No Known Activations