INDEX

Explanations

references to economic and human costs associated with errors and failures

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

Not in Any Lists

No Comments

Negative Logits

 $('[

-0.15

ulu

-0.14

enda

-0.14

imum

-0.14

olith

-0.14

 sele

-0.13

_WRAP

-0.13

LError

-0.13

 abuse

-0.13

Äĥ

-0.13

POSITIVE LOGITS

loss

0.17

alama

0.16

Loss

0.15

 loss

0.15

aight

0.15

alsa

0.14

ibase

0.14

 LOSS

0.14

 embarrassment

0.14

astreet

0.14

Activations Density 0.144%

references to economic and human costs associated with errors and failures

No Comments

No Known Activations