INDEX

Explanations

phrases involving comparisons or considerations of different perspectives or roles

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a comparison or a condition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

kefeller

-0.74

èĪ

-0.70

æĹ

-0.67

usha

-0.63

æķ

-0.63

ãģł

-0.62

ãĥį

-0.62

çīĪ

-0.61

itton

-0.60

Via

-0.59

POSITIVE LOGITS

pires

1.35

pired

1.21

far

1.10

 opposed

1.06

pire

0.97

king

0.96

ylum

0.93

 soon

0.93

 long

0.88

 regards

0.85

Activations Density 0.119%

phrases involving comparisons or considerations of different perspectives or roles

phrases that indicate a comparison or a condition

No Comments

No Known Activations