INDEX

Explanations

words related to mastery or expertise

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "master" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

OPLE

-0.75

gm

-0.70

enegger

-0.68

AAF

-0.66

IDER

-0.66

vernment

-0.65

DEN

-0.64

 retirees

-0.64

jriwal

-0.64

psey

-0.63

POSITIVE LOGITS

pieces

1.20

mind

1.02

piece

0.91

fully

0.85

class

0.84

work

0.82

stroke

0.80

classes

0.80

slave

0.79

ayne

0.77

Activations Density 0.012%

words related to mastery or expertise

occurrences of the word "master" in various contexts

No Comments

No Known Activations