INDEX

Explanations

dates formatted with a year component

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

ungan

-0.16

finder

-0.15

uto

-0.14

urette

-0.14

zl

-0.14

pname

-0.14

-ext

-0.14

åĲ¦

-0.14

-0.13

 Levine

-0.13

POSITIVE LOGITS

Pow

0.14

liers

0.14

GUI

0.14

ymax

0.14

azon

0.14

ackbar

0.14

.PageSize

0.14

 Wend

0.14

istrovstvÃŃ

0.14

Activations Density 0.030%

dates formatted with a year component

No Comments

No Known Activations