INDEX

Explanations

mentions of helmets in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to helmets

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

 Dull

-0.71

 Bulletin

-0.69

 Radical

-0.68

 Pulitzer

-0.68

VD

-0.67

tery

-0.66

tle

-0.66

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.65

ABE

-0.65

 Relations

-0.64

POSITIVE LOGITS

 helmet

1.15

 helmets

1.14

 Helmet

0.97

 worn

0.96

 goggles

0.90

 wearer

0.83

 headset

0.77

 mask

0.75

 sunglasses

0.75

lain

0.70

Activations Density 0.010%

mentions of helmets in various contexts

references to helmets

No Comments

No Known Activations