INDEX

Explanations

phrases mentioning normalcy or deviation from it

oai_token-act-pair · gpt-3.5-turbo

references to the concept of normalcy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

haw

-0.76

âĵĺ

-0.76

NRS

-0.75

Spot

-0.73

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.72

lest

-0.72

 Theft

-0.70

åĤ

-0.70

 Starship

-0.69

è¦

-0.69

POSITIVE LOGITS

normal

1.04

 normal

1.03

cy

1.00

isation

0.99

ization

0.93

 Normal

0.92

ordinary

0.88

 partName

0.86

izes

0.85

 sized

0.84

Activations Density 0.009%

phrases mentioning normalcy or deviation from it

references to the concept of normalcy

No Comments

No Known Activations