INDEX

Explanations

political or military figures or activities

oai_token-act-pair · gpt-3.5-turbo

references to political operatives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

umph

-0.79

yards

-0.72

Bah

-0.71

xual

-0.71

rounded

-0.70

reath

-0.69

ð

-0.68

alone

-0.67

ranged

-0.67

enge

-0.66

POSITIVE LOGITS

 operative

1.23

 operatives

1.01

eering

0.84

 Emails

0.79

sters

0.74

eers

0.74

leground

0.72

 posing

0.70

 infiltrated

0.68

 assassin

0.67

Activations Density 0.007%

political or military figures or activities

references to political operatives

No Comments

No Known Activations