INDEX

Explanations

instances of the word "that" followed by other words or phrases

oai_token-act-pair · gpt-3.5-turbo

phrases expressing a sense of "one thing" or "the first thing" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

aciously

-0.76

iscover

-0.66

hip

-0.66

enegger

-0.65

ahime

-0.65

aced

-0.63

bill

-0.60

ãĤ´ãĥ³

-0.59

gur

-0.58

ãĤ©

-0.58

POSITIVE LOGITS

 distinguishes

1.24

 separates

1.17

 bothers

1.16

 happens

1.12

 prevents

1.05

 arises

1.04

 happened

1.03

 emerges

1.03

 occurs

1.02

 justifies

0.98

Activations Density 0.178%

instances of the word "that" followed by other words or phrases

phrases expressing a sense of "one thing" or "the first thing" in various contexts

No Comments

No Known Activations