INDEX

Explanations

phrases related to geographical locations, particularly mentioning states

oai_token-act-pair · gpt-3.5-turbo

references to geographic locations and associated trends or characteristics

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

èĢħ

-0.74

 Sapphire

-0.73

Äĩ

-0.70

 shorth

-0.68

ĸļ

-0.67

verson

-0.62

Glass

-0.60

Ã¨re

-0.60

Weak

-0.59

ãĤ£

-0.59

POSITIVE LOGITS

ISSION

0.78

ikuman

0.77

 volunt

0.72

Topic

0.68

 toile

0.67

 sights

0.65

 ante

0.65

 seams

0.64

 ILCS

0.64

otos

0.64

Activations Density 0.149%

phrases related to geographical locations, particularly mentioning states

references to geographic locations and associated trends or characteristics

No Comments

No Known Activations