INDEX

Explanations

phrases related to political figures and governmental actions

oai_token-act-pair · gpt-3.5-turbo

references to government officials and their actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 traject

-0.74

 capt

-0.69

AGE

-0.68

âĶĢâĶĢâĶĢâĶĢ

-0.67

 bondage

-0.67

ihadi

-0.66

 analytic

-0.66

Missing

-0.66

 selves

-0.64

rift

-0.62

POSITIVE LOGITS

 vetoed

1.31

 appoint

1.19

 veto

1.18

 presided

1.16

 announced

1.14

announced

1.10

 ordered

1.09

 appointed

1.09

 issued

1.06

 convened

1.06

Activations Density 0.414%

phrases related to political figures and governmental actions

references to government officials and their actions

No Comments

No Known Activations