INDEX

Explanations

instances where the word "some" is followed by a number

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate conditional or variable scenarios

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

oppable

-0.79

ifax

-0.73

ideo

-0.70

rogens

-0.68

redit

-0.67

oki

-0.67

âĢİ

-0.67

aque

-0.67

ailable

-0.67

ilant

-0.66

POSITIVE LOGITS

 guise

0.95

 cases

0.94

 respects

0.90

 workplaces

0.89

 manner

0.88

 areas

0.87

 regions

0.86

 contexts

0.86

 households

0.84

 jurisdictions

0.81

Activations Density 0.114%

instances where the word "some" is followed by a number

phrases that indicate conditional or variable scenarios

No Comments

No Known Activations