INDEX

Explanations

phrases indicating a comparison or evaluation based on a certain criteria

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

rouse

-0.80

enh

-0.74

Correct

-0.74

orem

-0.71

anasia

-0.70

okes

-0.69

oked

-0.69

arez

-0.69

terior

-0.68

ernal

-0.68

POSITIVE LOGITS

how

0.98

 recent

0.89

 previous

0.70

recent

0.69

 current

0.68

 similarities

0.67

 rumors

0.66

its

0.66

 what

0.65

 there

0.65

Activations Density 0.099%

phrases indicating a comparison or evaluation based on a certain criteria

No Comments

No Known Activations