INDEX

Explanations

proper nouns and organizations

oai_token-act-pair · gpt-3.5-turbo

references to specific organizations, groups, or entities

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

;;;;

-0.59

thood

-0.59

$.

-0.57

é¾įå

-0.56

]+

-0.55

emale

-0.54

*.

-0.53

/,

-0.53

farious

-0.53

+.

-0.52

POSITIVE LOGITS

has

0.59

 will

0.49

 announced

0.49

 continues

0.49

 wants

0.48

!--

0.47

 continued

0.47

 responded

0.47

 began

0.47

 stopped

0.46

Activations Density 1.301%

proper nouns and organizations

references to specific organizations, groups, or entities

No Comments

No Known Activations