INDEX

Explanations

kitchen-related items in the text

oai_token-act-pair · gpt-3.5-turbo

references to kitchens and related features or items

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

âĢ¢âĢ¢

-0.78

ullah

-0.71

 Trooper

-0.71

RAY

-0.70

ioxide

-0.68

idas

-0.67

inates

-0.67

 USAF

-0.67

gow

-0.66

inant

-0.66

POSITIVE LOGITS

 sink

0.95

 cabinets

0.91

washer

0.88

ware

0.86

nut

0.84

ette

0.82

 cooks

0.82

 kitchens

0.81

 appliance

0.80

 sinks

0.79

Activations Density 0.026%

kitchen-related items in the text

references to kitchens and related features or items

No Comments

No Known Activations