INDEX

Explanations

evaluative adjectives that express a range of sentiments, from positive to negative

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

à¥įà¤¤à¤ķ

-0.17

/Branch

-0.15

(~(

-0.15

ÄĻ

-0.14

anco

-0.14

 congrat

-0.14

idis

-0.14

ElementException

-0.13

lsi

-0.13

inez

-0.13

POSITIVE LOGITS

 considering

0.24

 since

0.19

 Considering

0.18

 fact

0.18

 because

0.16

ãĥ¼ãĥĲ

0.16

 Howe

0.15

Considering

0.14

for

0.14

Activations Density 0.147%

evaluative adjectives that express a range of sentiments, from positive to negative

No Comments

No Known Activations