INDEX

Explanations

the word "weak" in various contexts and intensities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "weak" or variations thereof

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ICAN

-0.82

APH

-0.74

ittee

-0.72

CENT

-0.72

 Hilton

-0.70

andise

-0.69

 Everest

-0.69

 Andromeda

-0.67

oration

-0.67

illion

-0.67

POSITIVE LOGITS

nesses

1.12

lings

1.11

 weak

0.91

ling

0.90

ening

0.87

est

0.87

ens

0.86

 weakest

0.86

ener

0.80

ly

0.78

Activations Density 0.010%

the word "weak" in various contexts and intensities

instances of the word "weak" or variations thereof

No Comments

No Known Activations