INDEX

Explanations

instances of the word "sub" and its variations, indicating a focus on "substitutions" or "subcategories."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

wort

-0.17

ader

-0.15

affe

-0.15

VERR

-0.15

aders

-0.15

fully

-0.14

igan

-0.14

wine

-0.14

ystone

-0.14

AGED

-0.14

POSITIVE LOGITS

scribers

0.41

sequent

0.41

stantial

0.40

stitution

0.39

sequently

0.38

scriptions

0.37

stitute

0.37

scribed

0.37

ordinate

0.36

stit

0.35

Activations Density 0.029%

instances of the word "sub" and its variations, indicating a focus on "substitutions" or "subcategories."

No Comments

No Known Activations