INDEX

Explanations

proper nouns such as names of movies, series, and people

oai_token-act-pair · gpt-3.5-turbo

references to animated films and television series

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

claw

-0.75

Protect

-0.73

 SERVICE

-0.72

afety

-0.72

 hairs

-0.68

Phys

-0.68

Hat

-0.67

upper

-0.66

odan

-0.66

reasonable

-0.65

POSITIVE LOGITS

 premiered

1.41

 starring

1.22

 sequels

1.16

 debuted

1.12

 anthology

1.12

itled

1.07

 debut

1.02

 acclaimed

1.01

 sequel

1.01

 premie

1.00

Activations Density 0.682%

proper nouns such as names of movies, series, and people

references to animated films and television series

No Comments

No Known Activations