INDEX

Explanations

the word "ent", likely part of the phrase "entertainment"

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L1R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.1.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

wich

-0.18

 Team

-0.15

esser

-0.14

 Harmony

-0.14

osti

-0.14

orney

-0.14

ypress

-0.14

ussian

-0.14

esModule

-0.14

gage

-0.14

POSITIVE LOGITS

rello

0.16

.Automation

0.15

 serpent

0.15

_tunnel

0.15

undo

0.14

AYOUT

0.14

 Crosby

0.14

ainless

0.14

 Ling

0.13

irst

0.13

Activations Density 0.000%

the word "ent", likely part of the phrase "entertainment"

No Comments

No Known Activations