INDEX

Explanations

abbreviations or titles followed by a numerical value and a statement

oai_token-act-pair · gpt-3.5-turbo

statements or sections that present factual information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Klux

-0.85

ctic

-0.76

isoft

-0.75

avorite

-0.70

prus

-0.68

avy

-0.67

yss

-0.67

xit

-0.66

itsch

-0.65

JV

-0.65

POSITIVE LOGITS

ually

1.18

orial

1.08

ional

1.08

oids

0.99

itious

0.95

ially

0.92

oid

0.89

ored

0.86

Fact

0.85

ual

0.85

Activations Density 0.027%

abbreviations or titles followed by a numerical value and a statement

statements or sections that present factual information

No Comments

No Known Activations