INDEX

Explanations

phrases related to using something, usually in a metaphorical sense

oai_token-act-pair · gpt-3.5-turbo

instances of the words "tap" and its various forms, indicating actions related to accessing or utilizing something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

gur

-0.68

 Bagg

-0.66

CAS

-0.62

ften

-0.60

pite

-0.60

birth

-0.60

 AMERICA

-0.59

 Explain

-0.59

 Advent

-0.59

eq

-0.59

POSITIVE LOGITS

tap

1.02

tap

0.93

 tapping

0.93

 taps

0.88

Tap

0.85

apon

0.81

 tapped

0.80

apy

0.78

pour

0.77

deck

0.76

Activations Density 0.011%

phrases related to using something, usually in a metaphorical sense

instances of the words "tap" and its various forms, indicating actions related to accessing or utilizing something

No Comments

No Known Activations