INDEX

Explanations

phrases related to requirements or necessities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating obligations or requirements

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

 facult

-0.61

 Says

-0.60

 cares

-0.58

 DOES

-0.56

doesn

-0.55

fitting

-0.53

asso

-0.53

 loves

-0.51

history

-0.51

case

-0.50

POSITIVE LOGITS

be

1.29

 been

1.02

Be

0.98

Be

0.95

be

0.95

BE

0.93

been

0.90

 being

0.83

 Been

0.76

Being

0.72

Activations Density 0.570%

phrases related to requirements or necessities

phrases indicating obligations or requirements

No Comments

No Known Activations