INDEX

Explanations

phrases related to conversations or interactions between individuals

oai_token-act-pair · gpt-3.5-turbo

phrases indicating assertiveness or strong opinions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

'."

-0.70

 Pelosi

-0.67

 Sinclair

-0.66

Pai

-0.65

NOW

-0.64

'>

-0.64

cms

-0.63

pering

-0.60

.'"

-0.59

»Ĵ

-0.59

POSITIVE LOGITS

 definitely

0.94

initely

0.88

laughs

0.87

Laughs

0.80

 Originally

0.78

wcsstore

0.77

 yeah

0.77

 depends

0.75

 absolutely

0.74

Originally

0.73

Activations Density 0.529%

phrases related to conversations or interactions between individuals

phrases indicating assertiveness or strong opinions

No Comments

No Known Activations