INDEX

Explanations

comma-separated lists of actions or items

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and their frequency

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

Ī

-0.69

gate

-0.65

Ŀ

-0.61

 Yemeni

-0.59

¬¼

-0.57

ilege

-0.56

 suits

-0.55

 Rice

-0.54

essee

-0.54

Viz

-0.54

POSITIVE LOGITS

pee

0.74

pless

0.71

while

0.67

walk

0.66

apa

0.64

but

0.64

Wars

0.64

depending

0.63

 albeit

0.62

 depending

0.62

Activations Density 0.465%

comma-separated lists of actions or items

punctuation marks and their frequency

No Comments

No Known Activations