INDEX

Explanations

phrases related to authority or oversight

oai_token-act-pair · gpt-3.5-turbo

repeated conjunctions or phrases emphasizing connections between ideas

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

DX

-0.87

Ĭ±

-0.76

BOOK

-0.73

shift

-0.73

NVIDIA

-0.73

Tes

-0.70

Ĥİ

-0.70

busters

-0.69

---------

-0.68

Times

-0.68

POSITIVE LOGITS

 consequently

1.22

 furthermore

1.14

 thus

1.08

 therefore

1.05

 moreover

1.03

 thence

1.03

 thereby

0.99

 assorted

0.98

 secondly

0.97

 hence

0.96

Activations Density 0.295%

phrases related to authority or oversight

repeated conjunctions or phrases emphasizing connections between ideas

No Comments

No Known Activations