INDEX

Explanations

words related to law enforcement and legal matters

oai_token-act-pair · gpt-3.5-turbo

words related to safety and security

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ĸļ

-0.85

 depths

-0.74

ById

-0.74

 ponds

-0.72

ngth

-0.70

\<

-0.68

ItemThumbnailImage

-0.67

 precincts

-0.65

 depth

-0.64

 wells

-0.63

POSITIVE LOGITS

agonist

0.88

imity

0.81

incial

0.79

zai

0.78

vernment

0.75

bably

0.74

ament

0.74

nder

0.72

asus

0.72

 hers

0.72

Activations Density 0.091%

words related to law enforcement and legal matters

words related to safety and security

No Comments

No Known Activations