INDEX

Explanations

expressions related to recognition or gratitude

oai_token-act-pair · gpt-3.5-turbo

expressions of gratitude and personal achievement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 omin

-0.78

NetMessage

-0.67

 approving

-0.67

pmwiki

-0.66

livious

-0.65

Requires

-0.65

idges

-0.62

 Lovecraft

-0.62

suggest

-0.59

 incomp

-0.59

POSITIVE LOGITS

 myself

1.28

my

1.00

Coach

0.88

 Coach

0.83

 sacrific

0.82

 teammates

0.82

 athlet

0.79

 audition

0.79

 coaching

0.77

 workouts

0.75

Activations Density 0.552%

expressions related to recognition or gratitude

expressions of gratitude and personal achievement

No Comments

No Known Activations