INDEX

Explanations

references to flags

oai_token-act-pair · gpt-3.5-turbo

references to flags, particularly in a political or cultural context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

nesota

-0.87

hower

-0.84

lihood

-0.81

theless

-0.81

STD

-0.78

 subsequ

-0.77

axter

-0.77

soever

-0.74

arios

-0.72

 Flavoring

-0.69

POSITIVE LOGITS

pole

1.16

rant

1.12

ging

1.12

staff

0.99

 bearer

0.91

 flags

0.88

 flag

0.87

Flag

0.86

inating

0.81

ged

0.81

Activations Density 0.008%

references to flags

references to flags, particularly in a political or cultural context

No Comments

No Known Activations