INDEX

Explanations

words related to being user-friendly

oai_token-act-pair · gpt-3.5-turbo

instances of the word "friendly" and its variations, suggesting a focus on positive social interactions and environments

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

ulet

-0.91

amen

-0.83

illon

-0.79

anish

-0.78

angles

-0.76

Bone

-0.76

skill

-0.75

grave

-0.74

acca

-0.73

aughter

-0.73

POSITIVE LOGITS

glers

0.91

 confines

0.80

friendly

0.80

 alternatives

0.78

é¾įå¥ĳå£«

0.75

liness

0.74

 friendly

0.72

 Friendly

0.71

 minded

0.70

 alternative

0.70

Activations Density 0.021%

words related to being user-friendly

instances of the word "friendly" and its variations, suggesting a focus on positive social interactions and environments

No Comments

No Known Activations