INDEX

Explanations

repeated expressions of preference or affection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

ista

-0.17

/by

-0.16

 behalf

-0.16

ItemType

-0.16

idth

-0.15

ils

-0.15

ught

-0.14

uelles

-0.14

sel

-0.14

acco

-0.14

POSITIVE LOGITS

/dis

0.21

/lo

0.21

able

0.20

-minded

0.18

ably

0.17

elihood

0.16

 latter

0.15

Ike

0.15

WISE

0.15

to

0.15

Activations Density 0.048%

repeated expressions of preference or affection

No Comments

No Known Activations