INDEX

Explanations

negations or forms of denial in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

orp

-0.16

ndern

-0.16

avor

-0.15

avou

-0.15

kiem

-0.14

criptor

-0.14

hiba

-0.14

ãĤŃãĥ³ãĤ°

-0.14

thon

-0.14

Tan

-0.14

POSITIVE LOGITS

 thÃ¬

0.22

çļĦè¯Ŀ

0.22

 then

0.18

then

0.16

 THEN

0.16

åĪĻ

0.15

ï¼ĮåĪĻ

0.15

fte

0.15

Hak

0.15

olas

0.15

Activations Density 0.072%

negations or forms of denial in the text

No Comments

No Known Activations