INDEX

Explanations

phrases indicating sufficiency or insufficiency

oai_token-act-pair · gpt-3.5-turbo

phrases indicating sufficiency or adequacy 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ãĥĨ

-0.74

ãĥ¼ãĥĨãĤ£

-0.69

ãĥ¬

-0.65

ãĥĲ

-0.64

æ©Ł

-0.60

 Kend

-0.60

PHOTOS

-0.58

å°Ĩ

-0.58

ramid

-0.57

ä¿

-0.57

POSITIVE LOGITS

 comfortably

0.90

 worthwhile

0.84

 reliably

0.80

 oneself

0.79

 safely

0.70

 warrants

0.70

itant

0.70

paren

0.70

 warrant

0.68

 anymore

0.68

Activations Density 0.298%

phrases indicating sufficiency or insufficiency

phrases indicating sufficiency or adequacy in various contexts

No Comments

No Known Activations