INDEX

Explanations

references to inner aspects or workings

oai_token-act-pair · gpt-3.5-turbo

references to "inner" aspects or qualities related to various topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orthy

-0.85

atoes

-0.81

essors

-0.81

oulos

-0.77

HAM

-0.77

eday

-0.77

netflix

-0.76

enegger

-0.75

DragonMagazine

-0.74

NRS

-0.74

POSITIVE LOGITS

 workings

1.19

most

1.05

 inner

0.83

 circle

0.80

 sanct

0.79

 circumference

0.78

 Inner

0.75

neath

0.75

ranean

0.75

 thigh

0.73

Activations Density 0.008%

references to inner aspects or workings

references to "inner" aspects or qualities related to various topics

No Comments

No Known Activations