INDEX

Explanations

numeric values indicating an increase or decrease in quantity or quality

oai_token-act-pair · gpt-3.5-turbo

phrases indicating an increase in various metrics or prices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.66

Maker

-0.65

pers

-0.64

________________________

-0.64

ooo

-0.63

 Savior

-0.62

Closure

-0.62

 Allow

-0.62

 Aliens

-0.61

Â¯Â¯Â¯Â¯

-0.61

POSITIVE LOGITS

graded

1.01

xual

0.93

grading

0.92

grades

0.91

acebook

0.89

river

0.87

stairs

0.85

gradation

0.84

stage

0.78

ollen

0.77

Activations Density 0.049%

numeric values indicating an increase or decrease in quantity or quality

phrases indicating an increase in various metrics or prices

No Comments

No Known Activations