INDEX

Explanations

mentions of differing opinions or perspectives

oai_token-act-pair · gpt-3.5-turbo

repeated references to the term "Others."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 both

-0.71

 just

-0.69

 basically

-0.67

 ticket

-0.66

 level

-0.64

 peak

-0.64

 pretty

-0.63

utra

-0.60

 every

-0.59

 attendance

-0.59

POSITIVE LOGITS

 Others

3.85

Others

2.72

 Other

1.90

 Another

1.43

Other

1.37

 others

1.36

 Many

1.25

 Someone

1.23

 Some

1.21

 Their

1.16

Activations Density 0.016%

mentions of differing opinions or perspectives

repeated references to the term "Others."

No Comments

No Known Activations