INDEX

Explanations

locations or proper nouns related to various countries or people

oai_token-act-pair · gpt-3.5-turbo

instances of specific terms or keywords, particularly related to geographical locations or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 PACs

-0.66

Marie

-0.62

 dime

-0.58

 Alban

-0.58

 Galile

-0.56

 Jury

-0.56

ppings

-0.55

 Fargo

-0.55

ources

-0.55

ngth

-0.55

POSITIVE LOGITS

istan

0.85

ooth

0.77

ĪĴ

0.71

ylon

0.68

Nur

0.67

acle

0.67

dit

0.66

hammad

0.66

nia

0.65

illus

0.65

Activations Density 0.297%

locations or proper nouns related to various countries or people

instances of specific terms or keywords, particularly related to geographical locations or entities

No Comments

No Known Activations