INDEX

Explanations

adverbs and adjectives related to political contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

erez

-0.16

able

-0.16

ables

-0.16

ORB

-0.16

ird

-0.15

elle

-0.15

edula

-0.15

stime

-0.15

 Voll

-0.14

ive

-0.14

POSITIVE LOGITS

 speaking

0.35

-speaking

0.29

 Speaking

0.28

 sound

0.26

sound

0.25

Speaking

0.24

 minded

0.23

spe

0.22

/math

0.22

 SOUND

0.22

Activations Density 0.057%

adverbs and adjectives related to political contexts

No Comments

No Known Activations