INDEX

Explanations

connecting words and phrases that indicate relationships or contrasts

oai_token-act-pair · gpt-3.5-turbo

transitional words or phrases indicating contrast and conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wcs

-0.65

seless

-0.62

window

-0.61

dozen

-0.61

vernment

-0.59

NRS

-0.59

fml

-0.59

-'

-0.58

ridor

-0.58

>)

-0.57

POSITIVE LOGITS

 although

0.83

 secondly

0.82

 hence

0.81

 luckily

0.81

 namely

0.79

 Extras

0.76

 Whilst

0.72

although

0.71

 furthermore

0.71

 unfortunately

0.70

Activations Density 0.701%

connecting words and phrases that indicate relationships or contrasts

transitional words or phrases indicating contrast and conditions

No Comments

No Known Activations