INDEX

Explanations

phrases related to historical events, specifically focused on World War II

oai_token-act-pair · gpt-3.5-turbo

references to World War II

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 snipp

-0.76

ciating

-0.74

Wiz

-0.74

Choice

-0.74

BOOK

-0.72

player

-0.70

heet

-0.70

votes

-0.69

Point

-0.68

name

-0.68

POSITIVE LOGITS

era

0.90

 vets

0.83

 Veterans

0.81

 bombers

0.80

 veterans

0.79

 Aviation

0.77

 Veteran

0.76

 Memorial

0.75

hower

0.74

 veteran

0.74

Activations Density 0.078%

phrases related to historical events, specifically focused on World War II

references to World War II

No Comments

No Known Activations