INDEX

Explanations

complex words related to specific concepts or theories

oai_token-act-pair · gpt-3.5-turbo

complex structures or frameworks across various 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

é¾įå

-0.83

 notor

-0.68

avorite

-0.67

Ire

-0.65

ãĥ¯ãĥ³

-0.65

ãĤĮ

-0.65

bledon

-0.64

ãģį

-0.63

ãĥĥãĥī

-0.63

 helicop

-0.61

POSITIVE LOGITS

 âĵĺ

0.65

 welcomes

0.54

 gestures

0.54

 aims

0.53

Â¶

0.53

 returns

0.52

↵↵

0.52

↵

0.48

 âĢº

0.47

 Posted

0.46

Activations Density 0.725%

complex words related to specific concepts or theories

complex structures or frameworks across various contexts

No Comments

No Known Activations