INDEX

Explanations

food menu items

oai_token-act-pair · gpt-3.5-turbo

references to various types of menus

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

lisher

-0.69

 Powers

-0.68

 Sidd

-0.67

soft

-0.67

UG

-0.64

 Judd

-0.63

 Johann

-0.63

founded

-0.62

 Contra

-0.62

axter

-0.61

POSITIVE LOGITS

isine

1.19

 menus

0.93

 meals

0.92

 selections

0.89

 choices

0.87

 menu

0.86

 salads

0.86

selection

0.82

 options

0.82

 restaurants

0.81

Activations Density 0.029%

food menu items

references to various types of menus

No Comments

No Known Activations