INDEX

Explanations

references to news publications and their editors

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

hea

-0.14

pher

-0.14

 Reserve

-0.14

.accept

-0.14

kin

-0.14

jer

-0.14

hle

-0.14

 Bond

-0.14

ren

-0.13

 Tort

-0.13

POSITIVE LOGITS

ypress

0.17

erus

0.16

anford

0.16

ÑĤÐµÑĢÐ½

0.15

ï»¿#

0.14

rylic

0.14

iÐ²

0.14

Ķ

0.14

-Sah

0.14

anka

0.14

Activations Density 0.060%

references to news publications and their editors

No Comments

No Known Activations