INDEX

Explanations

adjectives and descriptors indicating openness, consistency, and supportiveness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

oders

-0.16

nz

-0.15

ulations

-0.15

otation

-0.15

ulton

-0.15

lessness

-0.15

oplayer

-0.15

urance

-0.15

AGR

-0.14

chez

-0.14

POSITIVE LOGITS

 enough

0.29

 both

0.21

 BOTH

0.19

äºİ

0.18

both

0.18

 Enough

0.17

ily

0.17

Ã½v

0.17

itably

0.17

 bÃ¥de

0.16

Activations Density 0.159%

adjectives and descriptors indicating openness, consistency, and supportiveness

No Comments

No Known Activations

adjectives and descriptors indicating openness, consistency, and supportiveness

No Comments

No Known Activations