INDEX

Explanations

the presence of the word "unbeaten" and related terms indicating undefeated status

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

aces

-0.20

bur

-0.16

Divider

-0.16

bai

-0.15

ace

-0.15

SES

-0.14

zet

-0.14

ulet

-0.14

zas

-0.14

 Rosenberg

-0.14

POSITIVE LOGITS

LP

0.16

.nd

0.15

 Cater

0.15

 hammered

0.15

igo

0.15

NA

0.15

 Hammer

0.14

mal

0.14

µ¬

0.14

 hammer

0.14

Activations Density 0.037%

the presence of the word "unbeaten" and related terms indicating undefeated status

No Comments

No Known Activations