INDEX

Explanations

proper nouns and references to people or authors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

TRACT

-0.14

oupon

-0.14

çĨ

-0.14

ocz

-0.14

 incest

-0.13

oling

-0.13

/games

-0.13

imers

-0.13

ury

-0.13

 rival

-0.13

POSITIVE LOGITS

 Dank

0.16

uin

0.15

 Woodward

0.14

 Mour

0.13

ugin

0.13

 candle

0.13

*******************************************************************************/↵

0.13

uti

0.13

´

0.13

Activations Density 0.004%

proper nouns and references to people or authors

No Comments

No Known Activations