INDEX

Explanations

locations, specifically countries

oai_token-act-pair · gpt-3.5-turbo

country names and locations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bats

-0.77

cession

-0.67

changes

-0.67

tails

-0.64

 specials

-0.63

helps

-0.61

 signs

-0.61

paralle

-0.61

tain

-0.61

-0.60

POSITIVE LOGITS

;;;;;;;;;;;;

0.75

 Territory

0.72

tu

0.71

 Dept

0.71

nown

0.70

 Miko

0.69

atu

0.67

alkyrie

0.65

 Headquarters

0.65

anguage

0.62

Activations Density 0.066%

locations, specifically countries

country names and locations

No Comments

No Known Activations