INDEX

Explanations

expressions of regret and actions associated with mistakes or do-overs

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

hog

-0.23

hu

-0.23

ham

-0.23

hunt

-0.22

er

-0.22

hud

-0.22

halt

-0.21

hire

-0.21

house

-0.21

ity

-0.21

POSITIVE LOGITS

ting

0.91

ted

0.69

tings

0.66

ters

0.65

TING

0.52

ter

0.51

ten

0.50

tes

0.48

te

0.48

ta

0.47

Activations Density 0.113%

expressions of regret and actions associated with mistakes or do-overs

No Comments

No Known Activations