INDEX

Explanations

references to being new or newness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

erk

-0.16

urge

-0.15

Ø¨Ø¹

-0.14

edian

-0.14

zoom

-0.13

RegularExpression

-0.13

eso

-0.13

 sneak

-0.13

eck

-0.13

POSITIVE LOGITS

oje

0.17

aisy

0.16

Ø§ÙĨÙĩ

0.15

 Glasses

0.15

levator

0.15

Descriptors

0.14

APPER

0.14

 Chall

0.14

nite

0.14

Activations Density 0.009%

references to being new or newness

No Comments

No Known Activations