INDEX

Explanations

mentions of various islands

oai_token-act-pair · gpt-3.5-turbo

references to specific islands

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

umers

-0.87

cedes

-0.86

chel

-0.80

byter

-0.80

eger

-0.79

ources

-0.76

nyder

-0.75

riction

-0.74

perature

-0.72

ilde

-0.71

POSITIVE LOGITS

 Island

1.33

 Islands

1.04

 Islanders

1.02

 Heights

0.88

 Ferry

0.86

 Siege

0.86

 Breaker

0.83

 Royale

0.82

 Beach

0.82

 Meadows

0.80

Activations Density 0.008%

mentions of various islands

references to specific islands

No Comments

No Known Activations