INDEX

Explanations

adjectives and phrases that describe well-structured or well-composed elements

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

-0.15

etical

-0.15

etty

-0.15

ery

-0.15

arat

-0.15

ITE

-0.14

ayne

-0.14

chner

-0.13

wers

-0.13

ERY

-0.13

POSITIVE LOGITS

rawer

0.15

leck

0.15

ingleton

0.15

ÐµÑĢÐ³

0.14

rou

0.14

anford

0.14

.ai

0.13

icket

0.13

metatable

0.13

OLID

0.13

Activations Density 0.068%

adjectives and phrases that describe well-structured or well-composed elements

No Comments

No Known Activations