INDEX

Explanations

terms related to concepts of inversion or reversal

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

pack

-0.17

éĢĶ

-0.15

amu

-0.15

 ÙĪØ§ØŃ

-0.15

paths

-0.14

bl

-0.14

amo

-0.14

Ø°Ø±

-0.14

tel

-0.14

 curtain

-0.14

POSITIVE LOGITS

invert

0.21

-down

0.21

 invert

0.20

 upside

0.20

 flipped

0.19

 flip

0.17

 inversion

0.17

çĵ¶

0.17

 flipping

0.17

Ä±nca

0.16

Activations Density 0.010%

terms related to concepts of inversion or reversal

No Comments

No Known Activations