INDEX

Explanations

terms related to abuse and abstract concepts

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

abras

-0.16

Ð¾Ñħ

-0.16

isans

-0.15

backs

-0.15

elas

-0.15

ering

-0.15

sch

-0.14

abric

-0.14

igan

-0.14

iers

-0.14

POSITIVE LOGITS

OLUTE

0.26

querque

0.23

STRACT

0.21

 Dhabi

0.20

bie

0.20

igail

0.20

bing

0.19

antly

0.19

olutely

0.19

bers

0.19

Activations Density 0.046%

terms related to abuse and abstract concepts

No Comments

No Known Activations