INDEX

Explanations

elements of structured data or lists, particularly involving numbers and categories

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

er

-0.18

ers

-0.16

ity

-0.16

urn

-0.15

↵

-0.15

mie

-0.15

ist

-0.14

Äį

-0.14

ubb

-0.14

ritch

-0.14

POSITIVE LOGITS

igan

0.16

ehler

0.15

#ac

0.15

'gc

0.15

 AppleWebKit

0.14

conde

0.14

Ð½Ð°Ñħ

0.14

/gtest

0.14

-cols

0.14

 Marketable

0.13

Activations Density 0.188%

elements of structured data or lists, particularly involving numbers and categories

No Comments

No Known Activations