INDEX

Explanations

the word "from."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ully

-0.16

istik

-0.15

 saja

-0.15

(disposing

-0.15

.Interop

-0.15

Î¹ÎºÎ±

-0.15

_tC

-0.14

ëħĢ

-0.14

lah

-0.14

skirts

-0.14

POSITIVE LOGITS

/to

0.47

 scratch

0.34

 whence

0.34

/by

0.32

/about

0.32

 within

0.32

 across

0.31

/of

0.28

scratch

0.28

 abroad

0.27

Activations Density 0.327%

the word "from."

No Comments

No Known Activations