INDEX

Explanations

phrases indicating a lack of effectiveness or voids in various contexts

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

avan

-0.17

asers

-0.16

ÑĨÐµÑĢ

-0.16

aser

-0.15

strcasecmp

-0.15

uja

-0.15

awa

-0.15

Ø¯ÙĬ

-0.15

ê·¼

-0.14

Drawable

-0.14

POSITIVE LOGITS

iration

0.15

uci

0.15

otime

0.15

_PID

0.14

 childhood

0.14

STA

0.14

ifa

0.14

 margins

0.14

ignon

0.13

enna

0.13

Activations Density 0.005%

phrases indicating a lack of effectiveness or voids in various contexts

No Comments

No Known Activations