INDEX

Explanations

instances of the word "to" with varying frequencies of occurrence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

orate

-0.15

quam

-0.14

iyon

-0.14

Ð¾ÑĪ

-0.14

å¹²

-0.14

jez

-0.14

Slo

-0.14

aleigh

-0.13

gow

-0.13

 Ð±ÑĥÐ´ÑĤÐ¾

-0.13

POSITIVE LOGITS

 raised

0.21

 maneuver

0.19

 Raised

0.18

 spice

0.17

 hitch

0.17

tug

0.17

 satisfy

0.16

sdk

0.16

 vars

0.16

dam

0.16

Activations Density 0.009%

instances of the word "to" with varying frequencies of occurrence

No Comments

No Known Activations