INDEX

Explanations

references to positions or rankings in sequences or competitions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

rire

-0.17

uali

-0.16

SCII

-0.15

.LayoutStyle

-0.15

ÙĪØ§Ø±

-0.15

uae

-0.15

ocre

-0.15

uml

-0.14

Ð¾Ð³Ð¾Ð´

-0.14

wik

-0.14

POSITIVE LOGITS

 part

0.20

 portion

0.19

0.17

est

0.16

few

0.16

 three

0.16

portion

0.16

ment

0.16

0.15

Activations Density 0.055%

references to positions or rankings in sequences or competitions

No Comments

No Known Activations