INDEX

Explanations

phrases related to identity or identification

oai_token-act-pair · gpt-3.5-turbo

references to identity and identification

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zzo

-0.83

Ħ¢

-0.78

 Pigs

-0.75

ttes

-0.74

avorite

-0.72

hiba

-0.68

Ļ½

-0.67

scill

-0.66

uden

-0.65

hari

-0.64

POSITIVE LOGITS

ifiers

1.19

ifications

1.18

ifiable

1.15

ifier

1.15

ifying

1.13

ieth

1.05

ified

1.05

ification

1.03

itarian

1.00

ity

0.99

Activations Density 0.017%

phrases related to identity or identification

references to identity and identification

No Comments

No Known Activations