INDEX

Explanations

phrases containing the term "first-person"

oai_token-act-pair · gpt-3.5-turbo

phrases related to first-person narratives in contexts of crime and education

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

 Richards

-0.70

 Canaver

-0.68

atile

-0.66

stanbul

-0.66

ulhu

-0.64

 Deity

-0.62

 DRAG

-0.62

ãĤ¼ãĤ¦ãĤ¹

-0.62

ulkan

-0.62

 Shroud

-0.62

POSITIVE LOGITS

generation

1.15

degree

1.09

eyed

1.05

contact

1.03

level

1.02

party

1.01

class

0.99

chance

0.97

sized

0.96

serving

0.92

Activations Density 0.032%

phrases containing the term "first-person"

phrases related to first-person narratives in contexts of crime and education

No Comments

No Known Activations