INDEX

Explanations

sections of code or programming constructs in the document

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ombat

-0.18

ictim

-0.15

.cum

-0.15

izzo

-0.14

ois

-0.14

Ð½Ð¾Ð¶

-0.14

Ð¼Ð¿

-0.14

ucker

-0.14

emplace

-0.14

ì°¸

-0.13

POSITIVE LOGITS

ache

0.15

 Derrick

0.15

 HomeComponent

0.14

urga

0.13

 Bene

0.13

jev

0.13

 tslib

0.13

 Broad

0.13

-Requested

0.13

 CDDL

0.12

Activations Density 0.010%

sections of code or programming constructs in the document

No Known Activations