INDEX

Explanations

phrases related to activities or events

oai_token-act-pair · gpt-3.5-turbo

topics related to competition and performance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

catentry

-0.58

00200000

-0.58

WIND

-0.55

à¼

-0.52

cember

-0.51

();

-0.51

 Loan

-0.50

 Airbus

-0.50

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.49

ãĥ´

-0.49

POSITIVE LOGITS

instead

0.65

ensitive

0.61

esides

0.61

 anymore

0.59

 frivol

0.59

wart

0.55

 nowadays

0.54

 solely

0.54

 versus

0.53

 unnecessarily

0.52

Activations Density 1.764%

phrases related to activities or events

topics related to competition and performance

No Comments

No Known Activations