INDEX

Explanations

sequences of numbers or numerical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

ehler

-0.15

UNKNOWN

-0.14

 Kushner

-0.14

 Bowl

-0.14

usta

-0.14

asel

-0.13

åħĲ

-0.13

ÑĢÐ°Ð¹

-0.13

ust

-0.13

chn

-0.13

POSITIVE LOGITS

ATAB

0.15

ser

0.15

irl

0.14

gal

0.14

.showMessage

0.14

_mB

0.14

 imper

0.14

 Innoc

0.14

dings

0.14

icode

0.14

Activations Density 0.010%

sequences of numbers or numerical references

No Comments

No Known Activations