INDEX

Explanations

mentions of government organizations and their actions

oai_token-act-pair · gpt-3.5-turbo

possessive forms of the word "government."

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

pload

-0.74

DonaldTrump

-0.73

ozy

-0.72

ilst

-0.70

ESPN

-0.68

sson

-0.68

bucks

-0.68

oons

-0.67

ols

-0.67

hari

-0.67

POSITIVE LOGITS

own

1.22

 inability

0.96

 intervention

0.96

 response

0.93

 Office

0.91

 disapproval

0.91

 stance

0.89

 flagship

0.89

 deliberations

0.89

 displeasure

0.88

Activations Density 0.261%

mentions of government organizations and their actions

possessive forms of the word "government."

No Known Activations