INDEX

Explanations

numerical values, particularly those indicating quantities or counts

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

çıį

-0.16

ippo

-0.14

.policy

-0.14

($)

-0.14

-0.13

arde

-0.13

Hashtable

-0.13

ctype

-0.13

áº£

-0.13

ekk

-0.13

POSITIVE LOGITS

ocket

0.19

luk

0.14

egra

0.14

 riot

0.14

ë¶Ħ

0.14

InitialState

0.13

DefaultValue

0.13

tvrt

0.13

ê°ľ

0.13

 ADVISED

0.13

Activations Density 0.053%

numerical values, particularly those indicating quantities or counts

No Comments

No Known Activations