INDEX

Explanations

the word "ceasefire" at different strengths

oai_token-act-pair · gpt-3.5-turbo

mentions of ceasefire agreements and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Leary

-0.88

Lear

-0.88

Magikarp

-0.76

Americans

-0.72

apore

-0.72

yon

-0.72

Kin

-0.71

Club

-0.70

aum

-0.70

alph

-0.69

POSITIVE LOGITS

 ceasefire

1.33

 truce

1.19

 treaty

0.92

 treaties

0.90

 blockade

0.85

 hostilities

0.82

 peacefully

0.82

 standoff

0.82

 disarm

0.82

 footing

0.80

Activations Density 0.013%

the word "ceasefire" at different strengths

mentions of ceasefire agreements and related terms

No Comments

No Known Activations