INDEX

Explanations

phrases indicating progress or achievement

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

ingles

-0.15

ç¥¥

-0.14

izons

-0.14

 fights

-0.14

ena

-0.14

 thoughts

-0.14

ÙĪØ±Ø´

-0.14

 Thoughts

-0.14

POSITIVE LOGITS

 prepar

0.19

 preparation

0.19

 experiment

0.19

 warning

0.18

 experimental

0.17

 ì¤Ģë¹Ħ

0.17

 Warning

0.16

 prep

0.16

experiment

0.16

.prepare

0.16

Activations Density 0.007%

phrases indicating progress or achievement

No Comments

No Known Activations