INDEX

Explanations

references to locations, particularly cities and landmarks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

è¶Ĭ

-0.23

 cÃłng

-0.17

 cuanto

-0.15

ÑģÑĤÐ¸ÑĤ

-0.15

íĥĦ

-0.15

hausen

-0.15

ibase

-0.14

ĵåĲį

-0.14

pat

-0.14

POSITIVE LOGITS

 Netherlands

0.21

 Holland

0.21

 Hague

0.19

 Dutch

0.18

Twe

0.16

rani

0.15

NL

0.15

 Valk

0.15

Lim

0.15

Ð»Ð°Ð½Ð´

0.15

Activations Density 0.013%

references to locations, particularly cities and landmarks

No Comments

No Known Activations