INDEX

Explanations

instances of the word "cur" indicating a focus on control or restriction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

oit

-0.15

ãģ¡ãģ¯

-0.15

igh

-0.15

igid

-0.15

urance

-0.15

ed

-0.15

orange

-0.14

ole

-0.14

 Hermes

-0.14

 Bolton

-0.14

POSITIVE LOGITS

ricular

0.25

iously

0.25

ricula

0.25

tail

0.23

few

0.23

 rent

0.22

atorial

0.21

Rent

0.21

iosity

0.21

yÅ¡

0.21

Activations Density 0.009%

instances of the word "cur" indicating a focus on control or restriction

No Comments

No Known Activations