INDEX

Explanations

specific historical names and terms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 hung

-0.17

abin

-0.15

aceous

-0.14

urdy

-0.14

yst

-0.14

cov

-0.14

tul

-0.13

 coding

-0.13

ape

-0.13

phem

-0.13

POSITIVE LOGITS

bedo

0.17

iver

0.17

.mixin

0.15

shaw

0.14

ÅĻet

0.14

WithContext

0.14

 consect

0.13

EMS

0.13

utow

0.13

Activations Density 0.010%

specific historical names and terms

No Comments

No Known Activations