INDEX

Explanations

mentions of the word "drug"

oai_token-act-pair · gpt-3.5-turbo

references to rugs and rug-related contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Korean

-0.66

 Sparkle

-0.64

 Proposition

-0.63

 Petra

-0.62

 Highlands

-0.62

 Founders

-0.61

 Pebble

-0.61

 Enlightenment

-0.59

 Dresden

-0.59

 Peninsula

-0.59

POSITIVE LOGITS

rug

1.32

uay

1.05

ular

0.91

ules

0.86

uese

0.85

unda

0.85

ulent

0.82

ula

0.81

recy

0.81

hess

0.81

Activations Density 0.006%

mentions of the word "drug"

references to rugs and rug-related contexts

No Comments

No Known Activations