INDEX

Explanations

words or phrases written on various objects

oai_token-act-pair · gpt-3.5-turbo

significant headlines and phrases that indicate important messages or topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

unker

-0.66

jri

-0.63

SPONSORED

-0.61

astern

-0.59

hematic

-0.59

allery

-0.59

odcast

-0.59

ennes

-0.58

ockets

-0.57

bilt

-0.56

POSITIVE LOGITS

"#

1.71

"'

1.71

1.65

"(

1.62

1.55

".

1.52

"-

1.50

"@

1.50

"\

1.49

"+

1.47

Activations Density 0.451%

words or phrases written on various objects

significant headlines and phrases that indicate important messages or topics

No Comments

No Known Activations