INDEX

Explanations

occurrences of the word "every" and related counting terms, signaling patterns or recurring themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ipation

-0.17

ile

-0.16

æ½®

-0.15

iction

-0.15

king

-0.14

ser

-0.14

inner

-0.14

adesh

-0.14

iner

-0.14

ilian

-0.14

POSITIVE LOGITS

unge

0.15

ç¯Ģ

0.15

amente

0.15

ernals

0.14

encil

0.14

anik

0.14

opak

0.14

erif

0.14

Ð½Ð¾Ð¶

0.14

acin

0.14

Activations Density 0.040%

occurrences of the word "every" and related counting terms, signaling patterns or recurring themes

No Comments

No Known Activations