INDEX

Explanations

phrases related to offering help or assistance

oai_token-act-pair · gpt-3.5-turbo

phrases indicating support or assistance

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

 fault

-0.65

mart

-0.64

Fed

-0.64

Champ

-0.63

agues

-0.62

WARN

-0.60

mAh

-0.60

oos

-0.59

uld

-0.58

ames

-0.58

POSITIVE LOGITS

 regards

1.20

 regard

1.07

standing

0.98

stood

0.94

 impunity

0.88

draw

0.87

 dignity

0.82

 respect

0.78

drawn

0.78

holding

0.70

Activations Density 0.112%

phrases related to offering help or assistance

phrases indicating support or assistance

No Comments

No Known Activations