INDEX

Explanations

phrases containing the word "ese."

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations or places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

rost

-0.68

 Cumm

-0.68

Atlanta

-0.64

razil

-0.63

MAT

-0.62

SAN

-0.61

 Humanity

-0.61

Allen

-0.60

ITS

-0.59

Eva

-0.58

POSITIVE LOGITS

ese

1.31

esi

0.84

ktop

0.83

soever

0.82

hler

0.81

ecake

0.81

icter

0.80

efe

0.79

urry

0.78

qi

0.78

Activations Density 0.006%

phrases containing the word "ese."

references to geographical locations or places

No Comments

No Known Activations