INDEX

Explanations

punctuation and specific multi-character sequences in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

erno

-0.18

stin

-0.15

idal

-0.15

aken

-0.15

 Terminal

-0.14

 liter

-0.14

erner

-0.14

 gauche

-0.14

pare

-0.14

Ins

-0.14

POSITIVE LOGITS

meg

0.22

meg

0.16

 fitte

0.15

æİª

0.15

å¢ĥ

0.15

Backing

0.14

biz

0.14

isser

0.14

iny

0.13

uns

0.13

Activations Density 0.000%

punctuation and specific multi-character sequences in text

No Comments

No Known Activations