INDEX

Explanations

phrases related to popular culture references, particularly those involving fictional groups like X-Men

oai_token-act-pair · gpt-3.5-turbo

references to the "X-Men" franchise

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

heit

-0.72

ufact

-0.70

aturdays

-0.67

Cla

-0.63

 passage

-0.62

PDATE

-0.61

aceae

-0.60

lav

-0.60

leneck

-0.59

steen

-0.58

POSITIVE LOGITS

reme

0.92

modules

0.79

Large

0.77

hedral

0.72

Pacific

0.72

Fine

0.71

axis

0.70

axies

0.70

uni

0.70

Org

0.69

Activations Density 0.198%

phrases related to popular culture references, particularly those involving fictional groups like X-Men

references to the "X-Men" franchise

No Comments

No Known Activations