INDEX

Explanations

mentions of the word "veto"

oai_token-act-pair · gpt-3.5-turbo

references to the concept of veto power

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

arist

-0.76

Sea

-0.73

EAR

-0.73

 Feet

-0.71

ndra

-0.69

onductor

-0.68

istics

-0.67

annis

-0.66

nsic

-0.65

asma

-0.65

POSITIVE LOGITS

 veto

1.21

 vetoed

1.13

zai

0.93

 bluff

0.78

 unanimously

0.74

 amnesty

0.73

 filib

0.72

 override

0.72

 snowball

0.69

essee

0.69

Activations Density 0.018%

mentions of the word "veto"

references to the concept of veto power

No Comments

No Known Activations