INDEX

Explanations

proper nouns related to locations, such as plazas or streets

oai_token-act-pair · gpt-3.5-turbo

references to specific locations, particularly plazas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

liest

-0.85

nesses

-0.75

ERAL

-0.74

lessly

-0.71

loss

-0.71

bler

-0.70

manship

-0.67

gling

-0.66

matter

-0.66

ebook

-0.65

POSITIVE LOGITS

 Plaza

1.14

 Hotel

0.91

 plaza

0.90

aurus

0.88

 Hilton

0.83

adium

0.80

atl

0.79

 Blvd

0.78

itect

0.75

 Boulevard

0.73

Activations Density 0.016%

proper nouns related to locations, such as plazas or streets

references to specific locations, particularly plazas

No Comments

No Known Activations