INDEX

Explanations

references to governmental or national security entities

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

Negative Logits

prung

-0.16

å¼ĺ

-0.16

à¥Ģà¤¶

-0.15

áº¾

-0.15

ixe

-0.15

ParticleSystem

-0.14

ç¨

-0.14

unken

-0.14

aniel

-0.14

POSITIVE LOGITS

0.18

amer

0.15

0.14

 Dale

0.14

in

0.14

hei

0.14

alement

0.14

bone

0.14

 Warren

0.14

boom

0.14

Activations Density 0.027%

references to governmental or national security entities

No Known Activations