INDEX

Explanations

phrases related to specific names or entities, particularly the name "Perkins."

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and entities, particularly those related to Perkins

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

yang

-0.78

aries

-0.74

eer

-0.71

uel

-0.70

utable

-0.70

mson

-0.68

una

-0.68

 Kuwait

-0.67

uesday

-0.67

tan

-0.66

POSITIVE LOGITS

 Perkins

0.93

phia

0.83

ortun

0.81

odder

0.80

ervative

0.73

heid

0.73

nown

0.73

ippi

0.72

ucker

0.71

ciating

0.70

Activations Density 0.032%

phrases related to specific names or entities, particularly the name "Perkins."

references to specific individuals and entities, particularly those related to Perkins

No Comments

No Known Activations