INDEX

Explanations

references to blogs

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "blog."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Lauder

-0.75

BuyableInstoreAndOnline

-0.72

 evils

-0.64

inho

-0.62

 Baal

-0.61

odka

-0.60

 Scotia

-0.60

 chloride

-0.60

 overpowered

-0.59

Yin

-0.59

POSITIVE LOGITS

osphere

1.33

gers

1.27

ging

1.13

ged

1.13

ger

1.09

post

1.06

spot

1.01

gments

0.94

posts

0.93

 posts

0.92

Activations Density 0.035%

references to blogs

occurrences of the word "blog."

No Comments

No Known Activations