INDEX

Explanations

keywords related to unlocking achievements or features

oai_token-act-pair · gpt-3.5-turbo

words related to unlocking or accessing achievements and features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

odor

-0.85

 Correct

-0.69

 Relative

-0.68

cemic

-0.68

gone

-0.68

geist

-0.67

 Drawn

-0.64

 Abram

-0.63

oss

-0.63

wig

-0.62

POSITIVE LOGITS

 unlocking

1.32

 unlocks

1.24

 unlocked

1.20

 unlock

1.18

 Unlock

1.04

locks

0.87

locking

0.86

locked

0.81

CLASSIFIED

0.80

ESCO

0.77

Activations Density 0.008%

keywords related to unlocking achievements or features

words related to unlocking or accessing achievements and features

No Comments

No Known Activations