INDEX

Explanations

words related to negativity or unpleasantness

oai_token-act-pair · gpt-3.5-turbo

words related to the concept of "pleasing" or "pleasure."

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

ahime

-0.73

Reviewer

-0.72

Downloadha

-0.68

uador

-0.68

dayName

-0.67

quickShipAvailable

-0.66

DERR

-0.65

è¦ļéĨĴ

-0.64

rolet

-0.60

é¾

-0.59

POSITIVE LOGITS

ple

1.48

ples

1.03

plet

1.02

ply

0.95

veyard

0.92

rette

0.85

tsky

0.82

yond

0.81

ople

0.79

iciary

0.79

Activations Density 0.007%

words related to negativity or unpleasantness

words related to the concept of "pleasing" or "pleasure."

No Comments

No Known Activations