INDEX

Explanations

phrases that emphasize user responsibility or guidance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

imar

-0.17

yte

-0.16

isen

-0.15

imet

-0.15

isch

-0.14

bem

-0.14

setattr

-0.14

anger

-0.14

aus

-0.14

autoload

-0.13

POSITIVE LOGITS

uras

0.16

.opensource

0.14

Jab

0.14

ÅĤug

0.14

ActionCreators

0.14

croft

0.14

processors

0.14

ä¸ĸç´Ģ

0.14

 poly

0.13

ancell

0.13

Activations Density 0.001%

phrases that emphasize user responsibility or guidance

No Comments

No Known Activations