INDEX

Explanations

specific high-activation characters or symbols, possibly indicating a particular formatting or coding structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

áº·c

-0.15

 Kang

-0.14

 believer

-0.14

attachment

-0.14

OperationException

-0.14

iec

-0.13

Spinner

-0.13

.visualization

-0.13

rit

-0.13

uru

-0.13

POSITIVE LOGITS

 Correction

0.18

 correction

0.18

 assist

0.17

 assists

0.17

 Dest

0.16

 Assist

0.15

 Correct

0.15

 corrective

0.15

 Stable

0.15

 Equal

0.15

Activations Density 0.002%

specific high-activation characters or symbols, possibly indicating a particular formatting or coding structure

No Comments

No Known Activations