INDEX

Explanations

words related to channels, such as television channels or pathways

oai_token-act-pair · gpt-3.5-turbo

references to various types of channels

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ever

-0.65

berman

-0.64

 Hockey

-0.63

 Raiders

-0.62

hiro

-0.61

inals

-0.60

 Moons

-0.58

 Pulitzer

-0.58

 Dodgers

-0.58

ourke

-0.58

POSITIVE LOGITS

 channel

1.06

 channels

1.00

Channel

0.94

lain

0.92

icut

0.83

channel

0.79

 cannabin

0.78

akra

0.78

izoph

0.77

ombat

0.77

Activations Density 0.010%

words related to channels, such as television channels or pathways

references to various types of channels

No Comments

No Known Activations