INDEX

Explanations

phrases indicating temporal or historical context

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

elman

-0.16

 Universities

-0.15

idge

-0.15

pta

-0.14

ikon

-0.14

hik

-0.14

 gran

-0.14

etic

-0.14

 Colleges

-0.13

omba

-0.13

POSITIVE LOGITS

 Fence

0.15

Nug

0.15

 Guys

0.15

tright

0.15

loys

0.14

WWW

0.14

íģ

0.14

ULO

0.14

utenberg

0.14

wiki

0.14

Activations Density 0.001%

phrases indicating temporal or historical context

No Comments

No Known Activations