INDEX

Explanations

phrases related to rankings or positions

oai_token-act-pair · gpt-3.5-turbo

references to national attributes or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

anwhile

-0.77

 interf

-0.75

 photoc

-0.72

NEC

-0.70

 Mous

-0.67

Syd

-0.67

 encount

-0.65

 decomp

-0.65

 transc

-0.64

 predec

-0.63

POSITIVE LOGITS

Ļ

1.57

¬

1.41

ı

1.37

¡

1.35

¤

1.34

ª

1.32

Ĵ

1.31

ĸ

1.28

į

1.26

ħ

1.26

Activations Density 0.278%

phrases related to rankings or positions

references to national attributes or conditions

No Comments

No Known Activations