INDEX

Explanations

terms related to directions and positions, particularly vertical concepts

oai_token-act-pair · gpt-3.5-turbo

terms related to vertical and horizontal orientations or arrangements

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

REDACTED

-0.86

mberg

-0.76

utic

-0.73

ghazi

-0.72

bats

-0.72

hell

-0.70

email

-0.70

atories

-0.70

ãģ®éŃĶ

-0.70

ggies

-0.69

POSITIVE LOGITS

 axis

1.05

 stripes

1.00

 stabil

0.97

 scrolling

0.93

 orientation

0.90

 displacement

0.89

 takeoff

0.88

ity

0.87

 layout

0.85

 rect

0.84

Activations Density 0.040%

terms related to directions and positions, particularly vertical concepts

terms related to vertical and horizontal orientations or arrangements

No Comments

No Known Activations