INDEX

Explanations

phrases or names related to political figures and events

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals, particularly political figures

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

istical

-0.75

imates

-0.73

ORE

-0.65

uate

-0.65

 Grateful

-0.64

 Sorcerer

-0.63

 Metatron

-0.62

 Mirror

-0.62

CoC

-0.61

redit

-0.61

POSITIVE LOGITS

lette

0.97

 Rouse

0.93

 Rousse

0.91

stal

0.88

ff

0.87

lin

0.80

LB

0.77

cia

0.77

ĸļ

0.76

utics

0.76

Activations Density 0.009%

phrases or names related to political figures and events

mentions of specific individuals, particularly political figures

No Comments

No Known Activations