INDEX

Explanations

phrases related to conflicts or confrontations

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions or events that involve confrontation or conflict

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Governments

-0.59

 mosques

-0.59

 prisons

-0.57

doms

-0.56

+,

-0.54

 Countries

-0.54

 arsen

-0.53

cience

-0.52

 providers

-0.52

 âĢº

-0.52

POSITIVE LOGITS

 another

0.93

another

0.84

an

0.80

0.76

 some

0.73

 gunfire

0.68

something

0.67

handedly

0.67

romptu

0.66

TWO

0.66

Activations Density 0.415%

phrases related to conflicts or confrontations

phrases indicating actions or events that involve confrontation or conflict

No Comments

No Known Activations