INDEX

Explanations

statements starting with "You can't"

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "you can't" or variations thereof

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

çīĪ

-0.75

RAD

-0.74

 Palest

-0.71

enegger

-0.71

BuyableInstoreAndOnline

-0.70

 Sandwich

-0.66

 Gleaming

-0.65

 Circus

-0.63

 theaters

-0.62

 colonial

-0.61

POSITIVE LOGITS

£

1.02

Ĵ

0.96

lege

0.95

Ķ

0.95

«

0.95

ı

0.93

º

0.91

ĵ

0.91

ĳ

0.91

¨

0.90

Activations Density 0.115%

statements starting with "You can't"

instances of the phrase "you can't" or variations thereof

No Comments

No Known Activations