INDEX

Explanations

formal or structured elements related to a list or table of contents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

_FIN

-0.16

asics

-0.15

 coarse

-0.15

 Dome

-0.14

 Dispatch

-0.14

otty

-0.14

 covert

-0.14

.Raw

-0.14

omon

-0.14

ErrorMsg

-0.14

POSITIVE LOGITS

AMPLE

0.17

eward

0.16

 Ð¾ÑĤÐ»Ð¸

0.14

.scalablytyped

0.14

prene

0.14

oken

0.14

oden

0.14

_padding

0.14

arget

0.14

utoff

0.14

Activations Density 0.023%

formal or structured elements related to a list or table of contents

No Comments

No Known Activations