INDEX

Explanations

phrases related to contrasting or adding on to information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating rules or regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Mechdragon

-0.87

 Dian

-0.68

IOR

-0.65

Winged

-0.63

Thirty

-0.62

 blunt

-0.60

 Kund

-0.60

Enlarge

-0.60

 Rwanda

-0.59

ANK

-0.58

POSITIVE LOGITS

 ought

0.93

should

0.89

nery

0.89

 wont

0.87

 SHOULD

0.84

 shouldn

0.78

thens

0.77

still

0.76

ought

0.76

will

0.75

Activations Density 0.085%

phrases related to contrasting or adding on to information

phrases indicating rules or regulations

No Comments

No Known Activations