INDEX

Explanations

mentions of historical names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

ulp

-0.16

à¹ģà¸Ļ

-0.16

avar

-0.16

anova

-0.15

irst

-0.15

coop

-0.14

sled

-0.14

%C

-0.14

igh

-0.14

avit

-0.14

POSITIVE LOGITS

 consect

0.14

anitize

0.13

 emphas

0.13

 Pension

0.13

shal

0.13

nÄħ

0.13

bib

0.13

fuse

0.13

icens

0.13

Activations Density 0.198%

mentions of historical names and titles

No Comments

No Known Activations