INDEX

Explanations

adjectives and expressions of value or quality, often indicating emotional or subjective perceptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

them

-0.19

 them

-0.17

eux

-0.16

him

-0.16

odo

-0.15

onto

-0.15

ont

-0.15

why

-0.15

Them

-0.14

POSITIVE LOGITS

 they

0.24

soever

0.21

 Ð¾Ð½Ð¾

0.20

it

0.19

 Ð¾Ð½Ð¸

0.18

 Ð²Ð¾Ð½Ð¸

0.16

idious

0.16

 THEY

0.15

he

0.15

we

0.15

Activations Density 0.041%

adjectives and expressions of value or quality, often indicating emotional or subjective perceptions

No Comments

No Known Activations