INDEX

Explanations

phrases indicating an action that has not yet occurred or been completed

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

lies

-0.80

 Species

-0.68

ths

-0.66

çīĪ

-0.65

 Needs

-0.64

Must

-0.63

now

-0.62

Days

-0.62

still

-0.61

probably

-0.58

POSITIVE LOGITS

 adequately

1.05

 bothered

0.96

 satisf

0.93

 properly

0.93

 harmed

0.92

icable

0.90

epad

0.89

 anywhere

0.89

 formally

0.88

 prosecuted

0.88

Activations Density 0.074%

phrases indicating an action that has not yet occurred or been completed

No Comments

No Known Activations