INDEX

Explanations

specific character sequences that likely represent a different language or encoding format

oai_token-act-pair · gpt-3.5-turbo

numerical values and measurements in scientific contexts

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

 apost

-0.78

ardless

-0.74

 wardrobe

-0.73

 occas

-0.69

 Creed

-0.69

 patriotic

-0.69

Mog

-0.68

 patri

-0.68

ldom

-0.67

atism

-0.66

POSITIVE LOGITS

Researchers

1.35

Scientists

1.30

DNA

1.20

Explore

1.12

Science

1.05

Testing

1.02

MRI

1.02

RNA

1.00

ccording

0.98

Research

0.96

Activations Density 0.338%

specific character sequences that likely represent a different language or encoding format

numerical values and measurements in scientific contexts

No Comments

No Known Activations