INDEX

Explanations

references to a particular brand or product called "Flip"

oai_token-act-pair · gpt-3.5-turbo

references to a specific brand or product named "Flip."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hall

-0.76

 halluc

-0.75

 radioactive

-0.75

 covenant

-0.73

 incapac

-0.72

REDACTED

-0.72

uld

-0.71

 verte

-0.69

 expectancy

-0.68

 blood

-0.66

POSITIVE LOGITS

 Flip

3.99

 Snap

1.44

Dip

1.34

 Pocket

1.31

 Twist

1.26

 Swap

1.21

 Trip

1.18

Rip

1.17

Fit

1.13

Moe

1.10

Activations Density 0.030%

references to a particular brand or product called "Flip"

references to a specific brand or product named "Flip."

No Comments

No Known Activations