INDEX

Explanations

patterns or structures in tabular data

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

erson

-0.19

onder

-0.17

ipel

-0.15

ippo

-0.15

orra

-0.15

ughters

-0.14

 Nicholson

-0.14

essen

-0.13

prung

-0.13

POSITIVE LOGITS

ogs

0.15

zin

0.15

æ¨

0.15

zas

0.15

lsa

0.14

uble

0.14

ê¸

0.14

lap

0.14

itm

0.14

zda

0.13

Activations Density 0.006%

patterns or structures in tabular data

No Comments

No Known Activations