INDEX

Explanations

instances of phrases with a disruptive or contrasting tone

oai_token-act-pair · gpt-3.5-turbo

instances of expressions of incredulity or surprise

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

utive

-0.77

reated

-0.75

encia

-0.70

ļéĨĴ

-0.70

Unique

-0.69

ahu

-0.68

irie

-0.68

orem

-0.66

edom

-0.65

axis

-0.63

POSITIVE LOGITS

huh

1.23

eh

1.17

 though

1.14

but

1.13

 although

1.11

 especially

1.03

 haha

1.00

 albeit

0.98

 however

0.91

 frankly

0.91

Activations Density 0.346%

instances of phrases with a disruptive or contrasting tone

instances of expressions of incredulity or surprise

No Comments

No Known Activations