INDEX

Explanations

phrases related to intentions or purposes

oai_token-act-pair · gpt-3.5-turbo

phrases indicating intentions behind actions or objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

uesday

-0.61

 Solitaire

-0.59

 Flavoring

-0.58

anon

-0.57

natureconservancy

-0.57

laws

-0.56

reports

-0.53

urus

-0.52

 knots

-0.52

 Videos

-0.51

POSITIVE LOGITS

to

0.97

 primarily

0.86

 solely

0.85

 principally

0.81

 Parenthood

0.76

 purely

0.76

 specifically

0.75

to

0.73

for

0.72

 chiefly

0.71

Activations Density 0.051%

phrases related to intentions or purposes

phrases indicating intentions behind actions or objects

No Comments

No Known Activations