INDEX

Explanations

the word "always."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.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

era

-0.17

emens

-0.16

emer

-0.15

lemen

-0.15

levision

-0.15

ÐµÑģÐ°

-0.15

illery

-0.15

eer

-0.15

les

-0.14

dap

-0.14

POSITIVE LOGITS

igator

0.17

theless

0.15

ignment

0.15

uded

0.14

rb

0.14

/src

0.14

 luÃ´n

0.14

ready

0.14

uge

0.14

Activations Density 0.041%

the word "always."

No Comments

No Known Activations