INDEX

Explanations

threatening actions and statements

oai_token-act-pair · gpt-3.5-turbo

phrases related to threats or actions taken against individuals or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

quickShipAvailable

-0.79

////////////////

-0.76

abiding

-0.76

 Partnership

-0.74

kered

-0.74

reetings

-0.73

erning

-0.72

ĨĴ

-0.71

DragonMagazine

-0.68

eret

-0.68

POSITIVE LOGITS

sue

1.29

 boycott

1.19

 withhold

1.15

 retaliate

1.12

 veto

1.08

 derail

1.07

 unleash

1.05

 revoke

1.05

 kill

1.04

 destroy

1.02

Activations Density 0.120%

threatening actions and statements

phrases related to threats or actions taken against individuals or entities

No Comments

No Known Activations