INDEX

Explanations

mentions of watches

oai_token-act-pair · gpt-3.5-turbo

references to watches

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

Reviewer

-1.01

«ĺ

-0.78

istically

-0.76

EngineDebug

-0.72

ally

-0.68

bably

-0.67

eno

-0.67

 chemotherapy

-0.66

Interstitial

-0.65

 Mayhem

-0.62

POSITIVE LOGITS

tower

1.51

dogs

1.44

dog

1.40

maker

1.09

fulness

1.02

watch

1.01

faces

1.01

strap

0.96

points

0.95

list

0.94

Activations Density 0.043%

mentions of watches

references to watches

No Comments

No Known Activations