INDEX

Explanations

sequences of equal signs and dashes, indicating sections or breaks in content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 13-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L13R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.13.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

á»ĳi

-0.19

itary

-0.15

nd

-0.14

obra

-0.14

áº¯p

-0.14

azel

-0.14

ceptor

-0.14

zel

-0.14

enberg

-0.14

zilla

-0.13

POSITIVE LOGITS

ETY

0.15

à¥įà¤ªà¤¤

0.15

../../../

0.14

riangle

0.14

äºĭ

0.14

Î¿ÏħÎ½

0.14

oice

0.14

ulty

0.13

ItemAt

0.13

Activations Density 0.015%

sequences of equal signs and dashes, indicating sections or breaks in content

No Comments

No Known Activations