INDEX

Explanations

the word "accordingly" and its variations, indicating a focus on conditionality or action based on previous statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 2-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L2R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.2.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

Ãłn

-0.17

ku

-0.16

oulder

-0.16

ught

-0.15

flamm

-0.15

aku

-0.15

ify

-0.15

kÃ¶

-0.14

POSITIVE LOGITS

anlÄ±

0.16

adesh

0.14

gom

0.14

aÅŁ

0.13

 å¯Į

0.13

pcodes

0.13

onn

0.13

imeo

0.13

Lf

0.13

avaÅŁ

0.13

Activations Density 0.012%

the word "accordingly" and its variations, indicating a focus on conditionality or action based on previous statements

No Comments

No Known Activations