INDEX

Explanations

keywords related to historical events, particularly World War II

oai_token-act-pair · gpt-3.5-turbo

mentions of World War

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

obook

-0.93

ITY

-0.80

ITIES

-0.77

OGR

-0.74

*/(

-0.72

aminer

-0.71

romeda

-0.69

66666666

-0.69

uckland

-0.66

Choice

-0.66

POSITIVE LOGITS

lord

1.06

lords

1.03

riors

0.94

rior

0.92

fare

0.82

 Patton

0.79

bler

0.78

fighter

0.76

Osw

0.73

II

0.72

Activations Density 0.024%

keywords related to historical events, particularly World War II

mentions of World War

No Comments

No Known Activations