INDEX

Explanations

verbs indicating advice, recommendation, or possibility

oai_token-act-pair · gpt-3.5-turbo

modal verbs indicating possibility or speculation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kills

-0.61

ettle

-0.61

 Malaysia

-0.58

photos

-0.56

 Malays

-0.56

 Malaysian

-0.55

 Lion

-0.54

iful

-0.54

adh

-0.53

ophon

-0.53

POSITIVE LOGITS

 nevertheless

0.73

 nonetheless

0.67

pmwiki

0.65

ãĥĩãĤ£

0.64

 sugg

0.63

 hereby

0.62

 surely

0.61

ãĥĦ

0.59

rack

0.58

DragonMagazine

0.57

Activations Density 1.328%

verbs indicating advice, recommendation, or possibility

modal verbs indicating possibility or speculation

No Comments

No Known Activations