INDEX

Explanations

references to political and historical events involving investigation or accountability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 yesterday

-0.17

 Yesterday

-0.16

.zero

-0.16

zim

-0.16

currently

-0.15

ernaut

-0.15

 currently

-0.15

 Currently

-0.15

 MEDIA

-0.15

DTV

-0.14

POSITIVE LOGITS

Advertisement

0.17

 tavern

0.17

 lecture

0.16

 correspond

0.16

 subscriptions

0.16

accounts

0.15

 Lady

0.15

 accounts

0.15

 patent

0.15

 republican

0.15

Activations Density 0.548%

references to political and historical events involving investigation or accountability

No Comments

No Known Activations