INDEX

Explanations

words that reference people and their attributes or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

zew

-0.17

@nate

-0.15

ideo

-0.15

raid

-0.15

TeV

-0.14

.nano

-0.14

duino

-0.14

esi

-0.14

ju

-0.14

uzz

-0.14

POSITIVE LOGITS

erd

0.16

äºķ

0.16

 Jean

0.15

esper

0.15

den

0.15

cano

0.15

ding

0.15

è

0.15

cle

0.14

Cro

0.14

Activations Density 0.057%

words that reference people and their attributes or achievements

No Comments

No Known Activations