INDEX

Explanations

instances of potential danger or threats that are close in time or likely to happen soon

oai_token-act-pair · gpt-3.5-turbo

terms indicating an impending threat or danger

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

rooms

-0.86

lua

-0.85

girls

-0.84

yards

-0.81

pel

-0.76

verbs

-0.75

yer

-0.74

tek

-0.71

girl

-0.71

strap

-0.71

POSITIVE LOGITS

 doom

1.26

 demise

1.01

 imminent

0.98

 impending

0.98

 famine

0.96

 arrival

0.93

 danger

0.92

 threat

0.90

 threats

0.89

 endanger

0.86

Activations Density 0.015%

instances of potential danger or threats that are close in time or likely to happen soon

terms indicating an impending threat or danger

No Comments

No Known Activations