INDEX

Explanations

references to philosophical concepts and contradictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

Î³Îº

-0.16

ylon

-0.16

antasy

-0.15

kazy

-0.15

ãĥ«ãĥķ

-0.14

$GLOBALS

-0.14

chw

-0.14

plied

-0.14

uard

-0.14

_vlog

-0.13

POSITIVE LOGITS

 judging

0.48

 according

0.44

 based

0.42

 Based

0.36

according

0.36

based

0.34

jud

0.32

Based

0.31

Jud

0.30

Jud

0.30

Activations Density 0.092%

references to philosophical concepts and contradictions

No Comments

No Known Activations