INDEX

Explanations

proper nouns related to people or places, particularly those related to warfare or geopolitical events

oai_token-act-pair · gpt-3.5-turbo

references to war and military conflict

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

aminer

-0.86

sembly

-0.83

essee

-0.82

ħĭ

-0.72

Ħ¢

-0.71

İĭ

-0.71

URES

-0.71

ĸļ

-0.70

UCT

-0.68

icles

-0.67

POSITIVE LOGITS

rior

1.44

fare

1.43

riors

1.35

ring

1.13

lord

0.97

lords

0.94

war

0.93

vard

0.91

neys

0.91

ney

0.90

Activations Density 0.013%

proper nouns related to people or places, particularly those related to warfare or geopolitical events

references to war and military conflict

No Comments

No Known Activations