INDEX

Explanations

references to individuals' political roles and affiliations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ÑĤÑĢÐ¾

-0.16

kest

-0.14

visa

-0.14

ãĥ©ãĥĥãĤ¯

-0.14

-visible

-0.14

ights

-0.13

 Bounty

-0.13

 Ð¿Ð»Ð°ÑĤ

-0.13

ï»¿↵↵

-0.13

owell

-0.13

POSITIVE LOGITS

 Chef

0.20

 Refer

0.20

stell

0.20

Le

0.20

 Dire

0.19

Che

0.19

Sek

0.18

 refer

0.18

 chef

0.18

che

0.17

Activations Density 0.019%

references to individuals' political roles and affiliations

No Comments

No Known Activations