INDEX

Explanations

terms related to targeting specific entities or groups

oai_token-act-pair · gpt-3.5-turbo

instances of the word "targeting"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

batch

-0.81

mpeg

-0.69

birth

-0.68

sis

-0.67

gamer

-0.66

aut

-0.66

SourceFile

-0.66

shire

-0.66

vol

-0.65

FIN

-0.64

POSITIVE LOGITS

 targeting

1.02

 targets

0.94

eering

0.85

 targeted

0.84

eers

0.84

oided

0.77

 target

0.75

 specificity

0.72

ataka

0.71

 killings

0.71

Activations Density 0.014%

terms related to targeting specific entities or groups

instances of the word "targeting"

No Comments

No Known Activations