INDEX

Explanations

mentions of a specific length of time ago

oai_token-act-pair · gpt-3.5-turbo

references to products or items in a contextually significant manner

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

ATK

-0.69

 edits

-0.66

[+

-0.66

email

-0.65

End

-0.65

 inquiries

-0.65

Edit

-0.64

vine

-0.64

 Assistance

-0.64

reports

-0.62

POSITIVE LOGITS

rouse

1.23

uras

1.03

lot

1.01

sembly

0.98

 couple

0.97

few

0.96

cess

0.95

 handful

0.93

 bunch

0.91

pron

0.91

Activations Density 0.269%

mentions of a specific length of time ago

references to products or items in a contextually significant manner

No Comments

No Known Activations