INDEX

Explanations

references to safe practices and environments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

acha

-0.17

otte

-0.16

olls

-0.14

ÅĻÃŃd

-0.14

laÅŁtÄ±r

-0.14

gc

-0.14

ohen

-0.14

aison

-0.14

Ð¾ÑģÑĮ

-0.14

_edge

-0.13

POSITIVE LOGITS

 safe

0.22

.safe

0.22

 Safe

0.22

 passage

0.21

 haven

0.21

 Haven

0.21

AreaView

0.20

 harbor

0.20

hav

0.20

haven

0.20

Activations Density 0.019%

references to safe practices and environments

No Comments

No Known Activations