INDEX

Explanations

locations or geographical entities mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

mentions of countries

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

ancies

-0.84

ollen

-0.69

essler

-0.68

ith

-0.67

oway

-0.67

ctive

-0.67

Afee

-0.66

lihood

-0.66

inker

-0.66

IU

-0.65

POSITIVE LOGITS

 citiz

0.88

oslov

0.86

 Airlines

0.86

istan

0.83

 Lumpur

0.82

Kyr

0.78

aution

0.75

 Airways

0.75

zbek

0.75

 natives

0.74

Activations Density 0.021%

locations or geographical entities mentioned in the text

mentions of countries

No Comments

No Known Activations