INDEX

Explanations

sentences that prompt the reader to take action or engage with content

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate something was missed or not seen

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

eteenth

-0.71

Officers

-0.65

mination

-0.62

ettel

-0.62

kas

-0.62

Aim

-0.61

ospel

-0.60

kiss

-0.59

reth

-0.55

dq

-0.55

POSITIVE LOGITS

 anything

0.94

any

0.77

me

0.76

 spoilers

0.71

PW

0.69

anything

0.69

 Pastebin

0.69

 yourselves

0.67

ANY

0.65

 yourself

0.65

Activations Density 0.158%

sentences that prompt the reader to take action or engage with content

phrases that indicate something was missed or not seen

No Comments

No Known Activations