INDEX

Explanations

mentions of the Hilton hotel

oai_token-act-pair · gpt-3.5-turbo

mentions of the Hilton brand

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oshenko

-0.93

icians

-0.89

ror

-0.85

ician

-0.77

anguage

-0.76

rieg

-0.75

rations

-0.74

ife

-0.74

inational

-0.72

icio

-0.72

POSITIVE LOGITS

coh

0.83

 guiActiveUnfocused

0.78

ZI

0.76

 loft

0.71

 Virtue

0.70

 volleyball

0.67

 Hilton

0.66

0.65

 Melania

0.65

 Ivanka

0.64

Activations Density 0.024%

mentions of the Hilton hotel

mentions of the Hilton brand

No Comments

No Known Activations