INDEX

Explanations

quotations indicating strong negative emotions or reactions

oai_token-act-pair · gpt-3.5-turbo

quotations expressing various emotional responses

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

İĭ

-0.89

²¾

-0.83

»Ĵ

-0.80

¬¼

-0.79

·

-0.73

ĪĴ

-0.68

Ĭ

-0.67

eb

-0.67

£ı

-0.66

erity

-0.66

POSITIVE LOGITS

 reads

0.94

 says

0.93

said

0.87

 said

0.84

 according

0.83

 recalls

0.76

 referring

0.75

 replied

0.73

 remarked

0.72

 exclaimed

0.71

Activations Density 0.088%

quotations indicating strong negative emotions or reactions

quotations expressing various emotional responses

No Comments

No Known Activations