INDEX

Explanations

mentions of specific policies or procedures that are in place

oai_token-act-pair · gpt-3.5-turbo

references to established systems or policies in operation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

anche

-0.68

ourge

-0.66

ĪĴ

-0.65

azes

-0.64

ostic

-0.63

Ranked

-0.62

iosity

-0.62

yssey

-0.62

ĳå£«

-0.62

lett

-0.61

POSITIVE LOGITS

 whereby

0.68

Aid

0.66

 preventing

0.65

 somewhere

0.64

for

0.63

 today

0.61

 financially

0.60

 defensively

0.60

 indefinitely

0.59

 barring

0.59

Activations Density 0.166%

mentions of specific policies or procedures that are in place

references to established systems or policies in operation

No Comments

No Known Activations