INDEX

Explanations

words related to wars and conflicts, particularly in Iraq and Afghanistan

oai_token-act-pair · gpt-3.5-turbo

titles or notable names associated with literary works or characters

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

hement

-0.82

wagen

-0.74

ersen

-0.74

 abolished

-0.72

uly

-0.70

 embr

-0.70

onge

-0.69

hene

-0.67

 adjustment

-0.67

 scrap

-0.67

POSITIVE LOGITS

Legendary

0.99

Creat

0.94

 Reloaded

0.90

Season

0.88

Episode

0.87

Volume

0.86

Advertisement

0.83

Magn

0.82

Hal

0.82

Associated

0.82

Activations Density 0.177%

words related to wars and conflicts, particularly in Iraq and Afghanistan

titles or notable names associated with literary works or characters

No Comments

No Known Activations