INDEX

Explanations

instances of tilde symbols, indicating some form of emphasis or annotation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.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

.LookAndFeel

-0.16

inge

-0.16

ãĥ¼ãĤ¹

-0.15

inged

-0.15

æľ

-0.15

 NONINFRINGEMENT

-0.15

efined

-0.14

orte

-0.14

sea

-0.14

ãĥ»

-0.14

POSITIVE LOGITS

ç·Ĵ

0.16

 vast

0.16

erin

0.14

Ø§ÙĦÙī

0.14

å¤ªéĥİ

0.14

olib

0.14

Î½Î¹

0.14

etÃŃ

0.14

gesi

0.14

argins

0.13

Activations Density 0.020%

instances of tilde symbols, indicating some form of emphasis or annotation

No Comments

No Known Activations