INDEX

Explanations

phrases indicating restrictions or permissions regarding the use of written material

oai_token-act-pair · gpt-3.5-turbo

phrases expressing the potential for permission or restrictions

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

 separating

-0.64

 separation

-0.63

 orchestrated

-0.61

 wielding

-0.59

 wages

-0.57

 oversaw

-0.57

 believing

-0.55

 robbing

-0.55

 forging

-0.55

 witnessing

-0.54

POSITIVE LOGITS

 vary

0.87

 differ

0.84

be

0.83

 expire

0.80

doi

0.76

Be

0.76

acio

0.75

 appear

0.75

BE

0.74

 contain

0.70

Activations Density 0.095%

phrases indicating restrictions or permissions regarding the use of written material

phrases expressing the potential for permission or restrictions

No Comments

No Known Activations